The distributed multiple-input multiple-output (MIMO) technique is one of the effective ways to mitigate the effect of atmospheric turbulence and reduce the packet error rate in free-space optical c ommunications.Because the time of laser propagation is different and random among atmospheric turbul ence links,signals arrive at the receiving antennas randomly and asynchronously,which causes the pr oblem of inter-symbol interference.But linear detection algorithms are more complicated to solve it f or the free-space optical (FSO) system with distributed transmit antennas.The valuable information obtaine d from signs of distributed receiving antennas depends on combining mode,and the performance of anti-interference is determined by the detection algorithm.Focusing on the problem above,this paper proposes the algorithm based on maximum likelihood detection and Neyman-Pearson (NP) fusion criterion for distributed receiving antennas system.Simulation results show that when receiv ing signal is synchronous, NP algorithm has better robustness,it is comparable to optimal com bining (OC) algorithm,and it is better than the equal gain combining (EGC) a lgorithm.When receiving signal is asynchronous, NP algorithm is better than EGC and OC algorithms,which improves spatial diversi ty performance in the distributed antennas system,suppressing inter-symbol interference.