The target tracking problem of bistatic MIMO radar under low SNR is studied

and a target tracking algorithm based on the improved AAJD (Adaptive Asymmetric Joint Diagonalization) is proposed. Firstly

the AAJD algorithm is improved to obtain the variable as the eigenvalue and the criterion of selecting the feature vector. The eigenvalue variables are used to find the eigenvectors corresponding to the large eigenvalue variables. And the problem of signal subspace expansion in AAJD algorithm is solved at low SNR. Secondly

the influence of the accumulation of the eigenvalue variables error is eliminated in the unsteady tracking state. The obtained signal subspace is more accurate. Since the estimated eigenvectors order is random at each time

the ESPRIT algorithm is improved to achieve the automatic pairing of transceiver angle of the same moment and the automatic association of the angle of the adjacent moment. The simulation results show that the improved AAJD algorithm can realize the angle tracking with low signal to noise ratio

and the convergence speed and stability performance are significantly better than AAJD algorithm.