In recent years, the statistical inference and algorithm for the complex diffusion process of incomplete data have become a hot topic that scholars of probability and statisticians are concerned with and calls for further study. Based upon the all-directional, multi-angle random dynamic information flow research, this article expands the information flow research in the ordinary information space to explore the broader, greater information measuring space. It discusses the recursive filtering algorithm of the optimal filtering and estimation of a kind of multi-dimensional, general part in the measurable, stable process in stochastic dynamic delivery system. The paper probes into and examines a series of valuable information in such generalized process. It produces the optimal recursive filtering equation and optimal estimation of random signals in this type of generalized process, thus provides a reliable theoretical foundation and an efficient mathematic method for further research of the optimal control of such a process and for further improvement of the efficiency of the sort of information transmission system.