Abstract: Space-time adaptive processing (STAP) is an effective method for moving target detection in airborne radar. However, detection of fast moving air dim target requires long accumulation time and hence range walk can occur. Keystone formatting combined with STAP provides a possible way for the above targets detection problem. However, due to serious Doppler ambiguity of the fast moving air dim targets, the direct use of Keystone formatting to the received data will degrade the STAP performance. Another problem is that Keystone formatting can’t compensate multiple targets’ range walk with different ambiguity numbers. In this paper, a new method combined STAP, Keystone formatting and Clean technique is proposed for multiple fast moving air dim targets detection. The new method removes clutter in the received data before Keystone formatting to avoid the performance degradation on STAP, and Clean technique is introduced for multiple target detection, by which one target’s range walk is compensated and then being detected based on the searched ambiguity number, after that, the data of that target is removed from the entire data and all the potential targets can be detected only by repeating the process. Effectiveness of the new method is verified via simulation examples.