Expert System has been grown so fast as a science that study how to make computer capable of solving problems that typically can only be solved by expert. It has been realized that the biggest challenge of developing expert system is the process include expert’s knowledge into the system. This research tries to model expert’s knowledge management using case based reasoning method. The knowledge itself is not inputted directly by the expert, but rather the system will learn the knowledge from what the expert did to the previous cases. This research takes image classification as the problem to be solved. As the knowledge development technique, we build decision tree by using C4.5 algorithm. Variables used for building the decision tree are the image visual features.