RSSI algorithm localization accuracy is very high in theory, but it is vulnerable to environmental interference and signal fluctuations, and if pursued all the details will greatly increase the complexity of the algorithm. This paper put forward a fuzzy localization algorithm against the weak points of RSSI algorithm in Wireless Sensor Networks,and applied to the RSSI location model. The new algorithm uses ridge-shaped distribution to fuzzification original RSSI signal. Then feedback and refine results according to the degree of approximation. The bigger value of fuzzy coefficient, the stronger anti disturbance ability, but the nodes have more or more similar lattice close degree, can not reflect the true state of the node, so the coefficient m should be based on attenuation index and environmental interference. Finally, simulation experiment shows that, the new algorithm can reduce the impact of environmental interference and signal fluctuations compared with RSSI and DV-Hop algorithm, which has high localization accuracy and stability.