The method of statistical analysis of data about aircraft engines operating times to removal from operation separately on each cause of failure is considered in this article. Such analysis allows to determinate the character of each failure. Initial (infant) failures, random (sudden) failures and degradation (deterioration) failures distinguish on character. The infant failures arise in initial period of engine operation and their intensity decreases in case of engine operating time increase. Random failures arise as a result of random external action (for example, in case of overloads), their intensity not depends from operating time. Intensity of degradation failures increases with increase of engine operating time. Character of failure is determined by means of Weibull distribution parameter аs. Estimation of Weibull distribution parameters is made by means of maximum likelihood method. If parameter аs ≈ 1, that failure is random and in order to this failure remove, it is necessary to increase engine structural strength and to improve their service conditions. If parameter estimation аs << 1, that this failure is infant and in order to such failure to prevent, it is necessary to increase efficiency of engine production, repair and assembly processes inspection. If parameter estimation аs >> 1, that this failure has degradation character and in order to such failure to prevent, it is necessary increase fatigue strength, erosion resistance and others properties of engine parts