Abstarct: Insurance companies offer different products, each of which can cover part of the risks, but unfortunately sometimes forgotten philosophy insurance and people with different tricks to "trying to cheat" and misappropriation of insurance "third party" are. "Fraud" of important issues losses to insurance companies and the insured is in all insurance fields. One way of identifying fraud "losses declared" the use of information fraud is discovered in the past. The data mining methods are widely used in discovering patterns in data. Using these methods can detect "damages fraud" in the "insurance industry" is useful. . In an experimental study these methods on real data, including information on the 1000 "case of damage" insurance third-party testing and performance of each method was used. The results showed that the accuracy of "algorithm" in education, 99.79% and 85.07%, which is in the testing of two basic algorithm is higher.