Electrical impedance tomography (EIT) is a visualized method for detecting the structural health of carbon-fiber-reinforced polymers (CFRPs). An EIT image reconstruction algorithm based on L1/L2 sparse regularization is proposed for underdetermination and ill-condition in EIT image reconstruction. In this method, the objective functional of the L1/L2 regularization term is constructed, a regularization parameter is added to modify the solution vector during the solution process, and a constraint interval is added in the iterative process to make the solution vector closer to the actual distribution. The simulation and experimental results show that compared with the conjugate gradient (CGLS), Tikhonov, and L1 regularization algorithms, the damage location and size reconstructed using the L1/L2 regularization algorithm are closer to the actual damage model, the damage identification is higher, and the electrode artifact is significantly improved. The proposed algorithm is a new method for applying EIT to the damage detection of CFRP laminates.