Camouflage netting is usually used against monitoring and reconnaissance, and SAR pictures of different bands and polarizations can provide synergy information of targets covered by the camouflage net. By qualitative theory analysis and quantitative data analysis, the effect of different bands and polarizations on targets under the net was researched. And the pixel and decision fusion algorithms were compared. Then a multi-bands SAR target detection method was provided based on decision fusion. Improved Neyman Pearson rule: regarding the probability of target detection and effective detection ratio as performance indexes, was used to select the most effective single-band or polarization resource. Moreover, scale filter was used to evaluate per target pixel and reduce the redundant information which came from single channel. The improved algorithm and the original one were evaluated by using effective detection ratio as performance index of fusion detection algorithm. Through the experimental results, we proved that the algorithm provided can reach high comprehensive performance of false rate-detection rate.