Preventive maintenance activities generate information that can help determine the causes of downtime and assist in setting maintenance schedules or alarm limits. When the amount of generated data becomes large, humans have difficulty understanding relationships between variables. In this thesis, we explore the applicability of data mining, a methodology for analyzing multi-dimensional datasets, to the maintenance domain, for find failures that cause event. Trough industrial partnerships, this project will demonstrate effectiveness of the proposed approach with actual Isfahan Regional Power Electric Company data. The objective is to extract and categorize machine components and subsystems and their associated failures using a novel approach that combines C-Fuzzy and genetic algorithm. With using this data mining method, we identify subsystems responsible for low overall equipment efficiency; recommend a preventive maintenance schedule and responses giving the most information about fault types. The data mining approach achieves good, easily understandable results within a short training time. The belief is that this will give the maintenance personnel a better understanding of when parts fail, allowing for s more accurate replacement schedule that could save money and improve safety.