In order to restore turbulence-degraded imaging of astronomical or space targets in real time, this paper proposes the multi-channel blind recognition method, which can be applied to the dynamic changes of atmospheric turbulence. Target imaging at different time after adaptive optical correction are regarded as multiple channels to establish the point spread function of system. The super-Laplace algorithm is used to solve the target after obtaining estimations of point spread function. Results show that there exists a mutual relationship between point spread functions at different moments, which satisfies the theory of multi-channel blind recognition. The Mean Square Error(MSE) between the solved point spread function and the original point spread function is in the order of 10^-30~10^-27 and the MSE between the recovered target image and the original target is in the order of 10^-5~10^-4. Research results provide a theoretical basis for real-time restoration of turbulence-degraded images.