This research presents a longan yield forecast model using data mining techniques using longan production data of longan farmers in Chiang Mai. Lamphun and Phayao Province, 215 cases and compare the performance of forecast models with data classification consists of techniques Tree algorithm J48, Hoeffding Tree, LMT, Randdom Tree and REP Tree. Rule technique using JRip algorithm, function technique using Multilayer Perceptron algorithm and correlation rule with Apriori algorithm. Model performance testing revealed that the Multilayer Perceptron algorithm was the most effective.