In the context of global climate change environment, China’s power grid is faced with many extreme weather challenges, especially the southern China power grid region, which faces typhoons, torrential rain, high temperature, drought, frost and other disasters that greatly affect the safe and stable operation of the power system and the normal social order in this region. This study proposes a risk assessment model combining a genetic algorithm-optimized neural network (GA-BP) with GIS spatial analysis to evaluate transmission line faults under extreme weather in southern China. Experimental results demonstrate the model’s effectiveness in identifying high-risk regions, with significant correlations between extreme precipitation, prolonged drought, and circuit failures.