The limited availability of electric vehicle (EV) charging load data in rural areas, due to low charging network coverage, poses a significant challenge for forecasting EV charging loads in rural tourism settings. Furthermore, existing studies mainly focus on short-term forecasting, with limited exploration of multi-timescale predictions. To address these issues, an enhanced multi-timescale charging load forecasting model for rural tourism EVs is proposed based on an improved Bayesian maximum entropy (BME) framework. First, an EV unit energy consumption model is established considering temperature and traffic impacts on EV travel behavior. Based on this, a short-term load forecasting model using the improved BME in developed. Next, by integrating an optimal grey model with tourist flow prediction, future stock of rural tourism EVs is forecasted, thereby deriving medium- and long-term charging load forecasts. Finally, simulation analysis using temperature and traffic data from a rural tourism area in Jiangsu Province is conducted to validate the effectiveness of the proposed method and predict the future development trend of rural tourism EV charging loads.