Models distributions of insurance claims is one of the main tasks of actuaries in insurance companies to estimate the price of insurance and loss reserves, fitting distribution of actual claims data is a problem closely related and not an easy task in Actuarial studies, mainly due to the nature of the claims data which characterized with a severe skew to the right. the research aims to use some of the probability distributions have the nature of skew positive, which is used widely in fitting insurance claims data, for example: (lognormal, weibull, fisk, lomax, logistic, paralogistic, inverse Gaussian, generalized extreme value distributions) and estimate the parameters of these distributions using the method of maximum likelihood to choose the best distribution of these distributions through several criteria (NLL, AIC, BIC).
The results showed that the generalized extreme value distribution is a better distribution fits insurance claims data which has less value BIC and AIC followed by fisk distribution followed by inverse Gaussian distribution. The researcher recommends to pricing comprehensive car insurance using generalized extreme value distribution and compare this pricing with the price used in the insurance company to come up with better pricing to determine a fair premium for the believers, enough to pay off the claims, achieve the margin profit to compete with other companies.