An　important　index　for　raw　milk　storage　in　a　dairy　farm　is　the　raw　milk　storage　temperature　which　directly　reflects　the　raw　milk　quality.　Meanwhile,　it＇s　hard　to　centrally　manage　numerous　dairy　farms　which　are　not　distributed　in　the　same　place.　We　aim　to　build　a　kind　of　raw　milk　monitoring　and　warning　equipment,　gateway　and　cloud　service　platform　to　solve　these　problems.　The　raw　milk　monitoring　and　warning　equipment　and　service　platform　were　designed　to　monitor　the　raw　milk　temperature　in　the　refrigerated　storage　tank　and　provide　a　warning　alarm　if　an　exception　occurred.　Data-driven　modeling　was　used　for　acquiring,　cleaning,　and　utilizing　data　to　solve　the　raw　milk　storage　problems.　The　raw　milk　monitoring　and　warning　management　system　provided　a　way　of　predicting　and　warning　for　raw　milk　storage　using　BP　Neural　Network　and　Fuzzy　Inference.　The　test　showed　that　the　BP　Neural　Network　and　Fuzzy　Inference　model　built　in　this　paper　had　a　good　performance　in　predicting　the　raw　milk　storage　temperature　and　reflecting　the　variation　of　raw　milk　temperature　in　raw　milk　storage　process.　The　platform　and　models　provided　a　method　to　manage　the　raw　milk　in　dairies　and　prevent　the　raw　milk　from　deteriorating　caused　by　the　rising　temperature.　©　2018　China　Agricultural　University