In this article the neurofuzzy systems of nonlinear filtering are offered. An example of the object movement in spherical system of coordinates is reviewed with the absence of its movements in the vertical plane. The state vector of the object is estimated by using nonlinear measurements of polar range and azimuth coordinates made by the two-coordinate radar station. The possibility of essential increase of filtering algorithms speed is established by using as the subtracting and fuzzy clustering and decomposition on the basis of three basic structures.