The accuracy of power load forecasting results is of great significance to the safe and stable operation of a power system. Considering short-term load forecasting under the influence of multiple meteorological factors, a new short-term power load forecasting method based on an improved Apriori correlation analysis and moth-flame optimization is proposed. First, an improved Apriori algorithm is used to analyze the correlation between meteorological factors and load. Unnecessary meteorological factors are removed from the results, and a human comfort evaluation index is introduced. Secondly, the reduced meteorological data combined with regional load data are used as input to the model. Finally, the short-term load forecasting model is established based on a long short-term memory neural network, and the global optimization model is combined with the moth-flame optimization algorithm. The results show that the load forecasting model can effectively improve the short-term load forecasting performance of a regional power grid.