Abstract
Nowadays Machinery Diagnostic becomes a major part for many industrial applications. It allows to predict and prevent of breakages. An analysis of the trends in the development of power machines show that the most advanced installations can be created using gas turbine technologies. Quite justified, many energy specialists consider the XXI century - the century of gas turbine technologies. It is very important to prevent gas turbine failure. In this paper investigated machine learning classification techniques with further implementation for fault detection in gas turbine running data trends. Investigation was done for real gas compression station running parameters.