Infrared small target detection under complex cloud background is a problem to be solved in the infrareddetection system. The key of the problem is about the distinguishment between the infrared small target and the cloudclutters. The traditional features that are used to differentiate them are closely collected with the absolute gray intensity，which is severely influenced by the change of background，so that，they cannot adapt the detection of infrared small targetunder complex cloud background. To distinguish the infrared small target and the cloud clutters efficiently，a group of imagefeature，constructing with local contrast features and segmented shape features is proposed. The feature group incorporatesthe advantages of both the two kinds of features，and fully expresses the difference of gray contrast and distinction of shape.Experimental results demonstrate that the proposed feature group can distinguish the infrared small target and the cloudclutters with much more efficiency.