对用BP网络进行逐时空调负荷预报的建模方法进行了研究。提出了两种类型负荷预报模型，基于建筑模型辩识的负荷预报法和基于时间序列的负荷预报法。并分四种具体情况进行了分析，得出了一种较为适合的负荷预报方法，这对于营冷空调系统优化控制的实施具有重要的意义。 A method for modeling the hourly air conditioning load forecasting using the BP network was studied. Two types of load forecasting models are proposed, load forecasting methods based on building model identification and load forecasting methods based on time series. The analysis was made in four specific situations and a more suitable method for load forecasting was obtained. This has important significance for the implementation of optimal control of camp cooling air conditioning system.