In the synthetic aperture radio telescope, the reconstruction of the radiation signal from the measured visibility function is an ill-posed inverse problem. Although compressed sensing technology has been successfully applied in synthetic aperture radio telescope imaging, the traditional compressed sensing algorithm uses L
1 norm to approximately replace L
0 norm, which brings some bias. To address this problem, a new imaging method of synthetic aperture radio telescope based on min-max concave penalty is proposed. The method uses the min-max concave penalty to approximate the L
0 norm and the fast iterative shrinkage-thresholding algorithm to solve the minimization model. In the iterative process, the regularization parameter is selected adaptively by using maximum likelihood estimation, and the convergence speed is improved by using restart and adaptive strategies. The experimental results show that the proposed method is superior to the current typical compressed sensing algorithms in terms of reconstruction accuracy and noise suppression, which proves its effectiveness.