The arresting plan of drug offender in community is a plan of police stations that needs skill and experience of polices in arresting. However, the
summary plan is used a lot of data for arrested drug offenders. The objective of this research is purposed factors analysis model of drug offender affecting drug crime using data mining techniques to help arresting plan of Nongbaulamphu police station overall 340 persons. This research compared 2 models in this experiment that are decision tress and naïve basian learning. The models can predicable factors for arresting plans of drug offenders for 3 types of drugs. The experimental result found that decision tress model gives accuracy to 91.18% which is higher than naïve baysian learning model. Keyword