To solve the problem of reconstructing complex sparse signals from linear measurements of additive white Gaussian noise (AWGN) under the condition of unknown sparse signal distribution, this paper proposes a fast adaptive complex approximate message passing (CAMP) algorithm with significantly reduced mean square error and adaptive to parameter selection. First, we establish a complex sparse signal distribution model. Secondly, the unknown parameters of the distribution model are estimated in each iteration, and the estimated values are used as prior information to obtain the complex shrinkage function with minimum mean square error (MMSE). Finally, an improved fast adaptive CAMP algorithm is obtained by combining the updated shrinkage function with the CAMP. The algorithm has the advantages of fast convergence, low computational complexity, small mean square error, and good robustness. Theoretical analysis and simulation verify the effectiveness of the proposed algorithm.