Necessary and sufficient conditions for almost sure asymptotic stability of solutions of stochastic dynamical systems generated by linear and nonlinear, nonautonomous ordinary sto-driven by square-integrable independent random variables (ξ n+1 ) n∈N with uniformly bounded quantities σ n ξ n+1 are in the center of this presentation. All conditions are explicitly expressed in terms of the coefficients α n , σ n , f and g. Kolmogorov's variant of the strong law of large numbers as well as martingale convergence and martingale representation theorems are applied to prove related results.