Based on distributed radar detection theory, this paper discusses the detection performance of multistatic radar with computer simulation technique. The detection contents include four kinds of well-known Swerling fluctuation radar returns. Various parameters such as decision strategy of the central processor and spatial distribution of input signal-to-noise ratios of the peripheral receivers can be put in as required. Four kinds of well-known Swerling fluctuation radar returns mixed with IF bandwidth limited Gaussian noise are modeled for simulation in the computer. The whole system detection simulated software is drawn up. The computer simulation program of combined detection of the five multistatic radars have been run on computer IBM-4381. The results show that on the premise of the false alarm probability of 10-4, it is possible to find probability in seventy five percent if majority decision is adopted. Therefore the improvement that the signal-noise ratio in this detection is 3dB higher than that in monostatic radar detection is achieved.