In this paper, the author analyzed an applicability of selected types of time-frequency distributions that belong to Cohen’s class and their reassignments for signals similar to those obtained during machinery diagnostics. At the first step of performed studies a synthetic multicomponent signal that contains both stationary and non-stationary components was analyzed using algorithms based on various time-frequency distributions. This allows for evaluating effectiveness of identification of particular components by applied time-frequency distributions and selecting a group of the most effective algorithms. At the second step, the selected time-frequency distributions were applied for analysis of signals acquired during diagnosis of rolling bearings in order to verify the effectiveness of identification of components responsible for a priori known faults occurred in bearings.