The solution of many quantitative problems is usually made while resolving the «intricacy» of initial database. If the initial database is vague, with the solutions being unstable as regards small variations of the initial database, then we have a trajectory divergence, characterized by the Lyapunov positive index. Existence of such an instability in the dynamic system is known to cause a chaotic solution behavior. In this way, the dynamic chaos is referred to non-correct problems according to Adamar. This paper proposes a new approach toward treatment of such systems which have random dynamics one way or another. The random dynamic system is made to correspond to such an equation set which is based on the regularization operator basic idea. The paper also demonstrates that this operator can be identified (during chaos description) with a fractional integrodifferential operator, the order of which is associated with the Lyapunov index or fractal dimensionality.