Now a days crime prevention is the challenging task for law enforcement bodies due to the huge volume of data for analysis (transform these huge amount of data to usable form). To overcome this problem, data mining application is the best solution that improves the performance of crime analysis for the purpose of proper resource allocation for crime prevention. The purpose of this study is to build a model using data mining techniques to investigate crime patterns focused on Bahir dar city administration. Therefore to achieve this objective decision tree and rule induction classification data mining techniques have been used in this study. The classification is done in to two cases with the target class names crime type and crime level. The experiment results show low performance which is not within the acceptable rage in both cases when we consider all crime types even though the model can identify crime prone areas in city administration and the criminal groups in specific crime types. However the model shows promising result on some crime types on the crime type target class. The cause of this poor performance of the model is the absence of enough predictor variables in the criminal record data base recorded by the police which needs further research by incorporating more attributes in the data base.