Аннотация:A general approach is studied to the problem of minimax-stochastically recurrent filtration of phase coordinates of multidimensional indefinitely-stochastic systems with discrete and continuous time in the presence of indirect measurements of not only phase coordinates proper but also certain components of the unknown input actions. The filtration algorithm is developed and some modifications and generalization of the algorithm are discussed. The results of numerical experiments are presented which allow to estimate the accuracy of obtained algorithms. It is concluded that the maximal accuracy is provided in the case when the control law is known.