With the rapid development of urbanization,the stable and efficient operation of the power system has become particularly important.This article aims to study a load dispatch strategy for urban power grids based on an improved genetic algorithm.Firstly,an overall framework for urban power grid load scheduling was established,which includes data collection and processing,load forecasting,optimized scheduling decisions,and scheduling execution and monitoring.Subsequently,improvements were made to the traditional genetic algorithm to enhance the search efficiency and quality of the solution.This article uses data from the PJM electricity market to conduct practical tests on the proposed strategy,verifying the application effect and actual performance of the improved genetic algorithm in power load dispatch.