To address the issues of load fluctuation and stability in the distribution network caused by unordered charging of electric vehicles (EVs) in residential communities, this paper proposes an EV charging scheduling strategy for residential communities based on Gaussian regression load forecasting and time-of-use (TOU) electricity pricing optimization. By analyzing the real data from a residential community, the electricity consumption and vehicle usage habits of users are identified. On this basis, a "same-day" forecasting approach is introduced, utilizing the historical data from the preceding two weeks as training data. Through the Gaussian regression prediction model, the base load of the residential community and the EV charging demand are predicted. Furthermore, integrating the TOU electricity pricing mechanism, a multi-objective optimization model aimed at minimizing the variance of the distribution network load and user charging costs is constructed. The particle swarm optimization (PSO) algorithm is employed to solve the objective function, and a comparative analysis of the distribution network performance under unordered and ordered charging conditions is conducted. The simulation results indicate that using data from the preceding two weeks can ensure that the prediction model is based on the latest load information while reducing the computational complexity of the model, thereby improving the accuracy and real-time performance of the predictions. Additionally, the proposed strategy in this paper not only reduces the load fluctuation and peak-to-trough difference in the distribution network but also decreases the charging costs for users, achieving the “peak shaving and valley filling” of the load curve.