Predicting　cooling　load　for　the　next　24　hours　is　essential　for　the　optimal　control　of　air-conditioning　systems　that　use　thermal　cool　storage.　This　study　investigated　modeling　methods　of　applying　the　general　regression　neural　network　(GRNN)　technology　to　predict　load.　The　single　stage　(SS)　and　double　stage　(DS)　prediction　methods　were　introduced.　Two　SS　and　two　DS　models　were　set　up　for　forecasting　the　next　24　hours＇　cooling　load.　Measured　data　collected　from　two　five　star　hotels　located　in　Sanya,　China,　were　used　to　train　and　test　these　models.　The　results　demonstrate　that　the　SS　method,　which　can　eliminate　the　necessity　for　measuring　and　predicting　meteorological　data,　is　much　simpler　and　reliable　for　predicting　the　cooling　load　in　practical　applications.