To address the problems of anisotropy and mode mixing in damage imaging on anisotropic composites laminates using ultrasonic Lamb waves, a sparse reconstruction imaging method based on S0 single-mode is proposed in this paper. In this approach, the sparsely arranged PZTs sensor group on the surface of the measured composite laminate is employed to excite and receive Lamb waves. The received signal shows sparse under a pre-built damage dictionary with damage scattering signals, and the damages can be local-ized and imaged using a sparse reconstruction method. Considering the velocity difference between S0 and A0 modes of received Lamb wave signal, the S0 mode is extracted by intercepting the signal before the peak of the first wave packet to eliminate the influence of Lamb wave mode mixing. In order to minimize the influence of anisotropy of composites laminates on Lamb wave propagation, the Lamb wave propagation direction is divided evenly into several sectors, and Lamb wave propagation direction within a certain sector is considered as propagation in a quasi-isotropic material. The sparse reconstruction is solved by using the BPDN algorithm, which can find the optimal solution with a small number of iterations in noise interference. The experiments on composite fiber laminate specimens show that the proposed method can accurately locate simulated single delamination damage and has less artifact interference compared to the DAS method.