This article presents the results of two experiments designed to study the effects of secondary motor task on the accuracy of time evaluation (for short spaces of time: 2, 3, 4 and 5 seconds). The participants learnt to hold a button during the aforementioned time intervals, while playing a computer game and getting feedback. The control group used the verbal counting strategy. The test subjects were offered two types of secondary motor task: the first group timed the button hold by counting the number of hand movements along their thigh; the second group estimated the necessary time span length by pressing the button while drawing circles and counting their quantity. The results indicated that using secondary motor task improved the accuracy of temporal interval production due to formation of more differentiated and stable subjective representations. A more complex secondary task (drawing circles on a piece of a paper) led to a better improvement in accuracy than a simpler task (moving hand along a thigh). The results are discussed in the context of Nikolay Bernstein’s idea of levels (hierarchy) of movement construction.