According to the problem of surface targets performances debasing under the condition of traditional constant false alarm rate (CFAR) inspection and low signal noise ratio, a new inspection method based on mean value and variance two-parameter is proposed. Firstly, target and background statistical characteristic difference maximum rule is used to estimate the mean value and variance of the inspected unit. And the final inspection image is obtained through multiplication on the mean value and variance of the inspected unit to obtain the statistical characteristics of the best mean value and variance statistical characteristic. And then, global CFAR inspection is used to inspect targets to guarantee a certain inspection probability and reduce false alarm probability. Simulation results show that the inspection characteristics of the method for low signal noise ratio surface target is better than that of traditional CFAR inspection.