Digital action mechanism of machining error propagation has been a hot research topic in recent years. A complicated machining system usually contains multiple stages. Basing on analyzing digital behaviors of machining process flow, a methodology of machining error monitoring and control is put forward, which is based on dynamic programming. Under this framework, state of machining feature is described with vector matrices, and then differential transition matrices are used to represent the influences of error sources on machining feature quality of workpiece. Basing on this, a general error propagation equation is derived. At last, an example of a three-stage machining processes is presented to illustrate the proposed methodology.