Modern aircraft have a large number of sensors that continuously record the performance of its various subsystems. In order to get the maximum benefit from the data obtained, they must be properly analyzed and interpreted. It is the developing field of predictive maintenance that allows you to monitor the condition of equipment, identify potential failures and prevent possible problems at an early stage. This work is devoted to predicting Remaining Useful Life (RUL) - the number of remaining operating cycles before system failure, based on historical data of engine operation. This problem can be formulated as a regression problem, and various machine learning methods were chosen to solve it. As a result of computational experiments, it was found that ensemble algorithms based on decision trees allow the most qualitative solution of the problem.