diff --git a/docs/user-guide/deployments.md b/docs/user-guide/deployments.md
index 45bf9f45c7..75e55a2761 100644
--- a/docs/user-guide/deployments.md
+++ b/docs/user-guide/deployments.md
@@ -402,7 +402,6 @@ at any time during the update is at most 130% of desired pods.
 minimum number of seconds for which a newly created pod should be ready
 without any of its containers crashing, for it to be considered available.
 This defaults to 0 (the pod will be considered available as soon as it is ready).
-__Note: This is not implemented yet__.
 
 ## Alternative to Deployments
 
diff --git a/pkg/controller/deployment/deployment_controller.go b/pkg/controller/deployment/deployment_controller.go
index ed0fc96225..6a99afcb74 100644
--- a/pkg/controller/deployment/deployment_controller.go
+++ b/pkg/controller/deployment/deployment_controller.go
@@ -208,8 +208,9 @@ func (d *DeploymentController) reconcileOldRCs(allRCs []*api.ReplicationControll
 	}
 	// Check if we can scale down.
 	minAvailable := deployment.Spec.Replicas - maxUnavailable
+	minReadySeconds := deployment.Spec.Strategy.RollingUpdate.MinReadySeconds
 	// Find the number of ready pods.
-	readyPodCount, err := deploymentutil.GetAvailablePodsForRCs(d.client, allRCs)
+	readyPodCount, err := deploymentutil.GetAvailablePodsForRCs(d.client, allRCs, minReadySeconds)
 	if err != nil {
 		return false, fmt.Errorf("could not find available pods: %v", err)
 	}
diff --git a/pkg/util/deployment/deployment.go b/pkg/util/deployment/deployment.go
index bcc0057f44..6d2e38d351 100644
--- a/pkg/util/deployment/deployment.go
+++ b/pkg/util/deployment/deployment.go
@@ -19,6 +19,7 @@ package deployment
 import (
 	"fmt"
 	"hash/adler32"
+	"time"
 
 	"k8s.io/kubernetes/pkg/api"
 	"k8s.io/kubernetes/pkg/api/unversioned"
@@ -132,19 +133,36 @@ func GetReplicaCountForRCs(replicationControllers []*api.ReplicationController)
 }
 
 // Returns the number of available pods corresponding to the given RCs.
-func GetAvailablePodsForRCs(c client.Interface, rcs []*api.ReplicationController) (int, error) {
-	// TODO: Use MinReadySeconds once https://github.com/kubernetes/kubernetes/pull/12894 is merged.
+func GetAvailablePodsForRCs(c client.Interface, rcs []*api.ReplicationController, minReadySeconds int) (int, error) {
 	allPods, err := getPodsForRCs(c, rcs)
 	if err != nil {
 		return 0, err
 	}
+	return getReadyPodsCount(allPods, minReadySeconds), nil
+}
+
+func getReadyPodsCount(pods []api.Pod, minReadySeconds int) int {
 	readyPodCount := 0
-	for _, pod := range allPods {
+	for _, pod := range pods {
 		if api.IsPodReady(&pod) {
-			readyPodCount++
+			// Check if we've passed minReadySeconds since LastTransitionTime
+			// If so, this pod is ready
+			for _, c := range pod.Status.Conditions {
+				// we only care about pod ready conditions
+				if c.Type == api.PodReady {
+					// 2 cases that this ready condition is valid (passed minReadySeconds, i.e. the pod is ready):
+					// 1. minReadySeconds <= 0
+					// 2. LastTransitionTime (is set) + minReadySeconds (>0) < current time
+					minReadySecondsDuration := time.Duration(minReadySeconds) * time.Second
+					if minReadySeconds <= 0 || !c.LastTransitionTime.IsZero() && c.LastTransitionTime.Add(minReadySecondsDuration).Before(time.Now()) {
+						readyPodCount++
+						break
+					}
+				}
+			}
 		}
 	}
-	return readyPodCount, nil
+	return readyPodCount
 }
 
 func getPodsForRCs(c client.Interface, replicationControllers []*api.ReplicationController) ([]api.Pod, error) {
diff --git a/pkg/util/deployment/deployment_test.go b/pkg/util/deployment/deployment_test.go
new file mode 100644
index 0000000000..54e044cd31
--- /dev/null
+++ b/pkg/util/deployment/deployment_test.go
@@ -0,0 +1,77 @@
+/*
+Copyright 2015 The Kubernetes Authors All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package deployment
+
+import (
+	"testing"
+	"time"
+
+	"k8s.io/kubernetes/pkg/api"
+	"k8s.io/kubernetes/pkg/api/unversioned"
+)
+
+func newPod(now time.Time, ready bool, beforeSec int) api.Pod {
+	conditionStatus := api.ConditionFalse
+	if ready {
+		conditionStatus = api.ConditionTrue
+	}
+	return api.Pod{
+		Status: api.PodStatus{
+			Conditions: []api.PodCondition{
+				{
+					Type:               api.PodReady,
+					LastTransitionTime: unversioned.NewTime(now.Add(-1 * time.Duration(beforeSec) * time.Second)),
+					Status:             conditionStatus,
+				},
+			},
+		},
+	}
+}
+
+func TestGetReadyPodsCount(t *testing.T) {
+	now := time.Now()
+	tests := []struct {
+		pods            []api.Pod
+		minReadySeconds int
+		expected        int
+	}{
+		{
+			[]api.Pod{
+				newPod(now, true, 0),
+				newPod(now, true, 2),
+				newPod(now, false, 1),
+			},
+			1,
+			1,
+		},
+		{
+			[]api.Pod{
+				newPod(now, true, 2),
+				newPod(now, true, 11),
+				newPod(now, true, 5),
+			},
+			10,
+			1,
+		},
+	}
+
+	for _, test := range tests {
+		if count := getReadyPodsCount(test.pods, test.minReadySeconds); count != test.expected {
+			t.Errorf("Pods = %#v, minReadySeconds = %d, expected %d, got %d", test.pods, test.minReadySeconds, test.expected, count)
+		}
+	}
+}
diff --git a/test/e2e/deployment.go b/test/e2e/deployment.go
index 8591504043..18a69a64c3 100644
--- a/test/e2e/deployment.go
+++ b/test/e2e/deployment.go
@@ -170,7 +170,7 @@ func testRollingUpdateDeployment(f *Framework) {
 		Expect(c.Deployments(ns).Delete(deploymentName, nil)).NotTo(HaveOccurred())
 	}()
 
-	err = waitForDeploymentStatus(c, ns, deploymentName, 3, 2, 4)
+	err = waitForDeploymentStatus(c, ns, deploymentName, 3, 2, 4, 0)
 	Expect(err).NotTo(HaveOccurred())
 }
 
@@ -251,7 +251,7 @@ func testRollingUpdateDeploymentEvents(f *Framework) {
 		Expect(c.Deployments(ns).Delete(deploymentName, nil)).NotTo(HaveOccurred())
 	}()
 
-	err = waitForDeploymentStatus(c, ns, deploymentName, 1, 0, 2)
+	err = waitForDeploymentStatus(c, ns, deploymentName, 1, 0, 2, 0)
 	Expect(err).NotTo(HaveOccurred())
 	// Verify that the pods were scaled up and down as expected. We use events to verify that.
 	deployment, err := c.Deployments(ns).Get(deploymentName)
diff --git a/test/e2e/util.go b/test/e2e/util.go
index 7a0e6a7eed..1f45aebd03 100644
--- a/test/e2e/util.go
+++ b/test/e2e/util.go
@@ -1843,7 +1843,7 @@ func waitForRCPodsGone(c *client.Client, rc *api.ReplicationController) error {
 
 // Waits for the deployment to reach desired state.
 // Returns an error if minAvailable or maxCreated is broken at any times.
-func waitForDeploymentStatus(c *client.Client, ns, deploymentName string, desiredUpdatedReplicas, minAvailable, maxCreated int) error {
+func waitForDeploymentStatus(c *client.Client, ns, deploymentName string, desiredUpdatedReplicas, minAvailable, maxCreated, minReadySeconds int) error {
 	return wait.Poll(poll, 5*time.Minute, func() (bool, error) {
 
 		deployment, err := c.Deployments(ns).Get(deploymentName)
@@ -1864,7 +1864,7 @@ func waitForDeploymentStatus(c *client.Client, ns, deploymentName string, desire
 		}
 		allRCs := append(oldRCs, newRC)
 		totalCreated := deploymentutil.GetReplicaCountForRCs(allRCs)
-		totalAvailable, err := deploymentutil.GetAvailablePodsForRCs(c, allRCs)
+		totalAvailable, err := deploymentutil.GetAvailablePodsForRCs(c, allRCs, minReadySeconds)
 		if err != nil {
 			return false, err
 		}