diff --git a/builtin/providers/aws/resource_aws_autoscaling_group.go b/builtin/providers/aws/resource_aws_autoscaling_group.go
index d5a87e33b57a..1fd52fa4e489 100644
--- a/builtin/providers/aws/resource_aws_autoscaling_group.go
+++ b/builtin/providers/aws/resource_aws_autoscaling_group.go
@@ -242,6 +242,7 @@ func resourceAwsAutoscalingGroupRead(d *schema.ResourceData, meta interface{}) e
 
 func resourceAwsAutoscalingGroupUpdate(d *schema.ResourceData, meta interface{}) error {
 	conn := meta.(*AWSClient).autoscalingconn
+	shouldWaitForCapacity := false
 
 	opts := autoscaling.UpdateAutoScalingGroupInput{
 		AutoScalingGroupName: aws.String(d.Id()),
@@ -253,6 +254,7 @@ func resourceAwsAutoscalingGroupUpdate(d *schema.ResourceData, meta interface{})
 
 	if d.HasChange("desired_capacity") {
 		opts.DesiredCapacity = aws.Int64(int64(d.Get("desired_capacity").(int)))
+		shouldWaitForCapacity = true
 	}
 
 	if d.HasChange("launch_configuration") {
@@ -261,6 +263,7 @@ func resourceAwsAutoscalingGroupUpdate(d *schema.ResourceData, meta interface{})
 
 	if d.HasChange("min_size") {
 		opts.MinSize = aws.Int64(int64(d.Get("min_size").(int)))
+		shouldWaitForCapacity = true
 	}
 
 	if d.HasChange("max_size") {
@@ -353,6 +356,10 @@ func resourceAwsAutoscalingGroupUpdate(d *schema.ResourceData, meta interface{})
 		}
 	}
 
+	if shouldWaitForCapacity {
+		waitForASGCapacity(d, meta)
+	}
+
 	return resourceAwsAutoscalingGroupRead(d, meta)
 }
 
diff --git a/website/source/docs/providers/aws/r/autoscaling_group.html.markdown b/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
index 6f2b0e51125c..c7cb07fa1afb 100644
--- a/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
+++ b/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
@@ -115,6 +115,11 @@ The first is default behavior. Terraform waits after ASG creation for
 `min_size` (or `desired_capacity`, if specified) healthy instances to show up
 in the ASG before continuing.
 
+If `min_size` or `desired_capacity` are increased in a subsequent update,
+Terraform will also wait for the correct number of healthy instances before
+continuing. Decreases to these numbers in updates do not currently activate any
+waiting behavior.
+
 Terraform considers an instance "healthy" when the ASG reports `HealthStatus:
 "Healthy"` and `LifecycleState: "InService"`. See the [AWS AutoScaling
 Docs](https://docs.aws.amazon.com/AutoScaling/latest/DeveloperGuide/AutoScalingGroupLifecycle.html)