A　fresh　air　system　with　solar　collectors　and　Latent　Heat　Thermal　Storage　(LHTS)　tank　filled　with　Phase　Change　Material　was　firstly　introduced　in　this　paper.　For　the　mathematical　model　of　the　fresh　air　system　is　difficult　to　accurately　establish,　the　adaptive　neuro-fuzzy　inference　controller　was　proposed　to　realize　the　parameter　identifications　of　the　system　applying　for　the　back-propagation　algorithm　of　BP　neural　network.　The　results　shows　that　the　modeling　and　parameter　identification　method　of　the　algorithm　using　ANFIS　can　approximate　the　characteristics　of　the　fresh　air　system　to　a　great　extent,　and　the　error　is　almost　zero.　©　2010　IEEE.