The paper proposes a method for evaluation of the structural and mechanical characteristics of the product material during machining. A process is considered in which the indicators of interest cannot be accurately determined without destroying the part, but with a certain degree of reliability can be established by indirect parameters. There is (in particular, according to the results of destructive tests) a connection between the known indirect and unknown (hidden) sought parameters, and in the most general form this connection is of an indeterminate nature. The known parameter is assumed to be continuous, the quantitative characteristic of the considered connection is the distribution function of the probability that the hidden parameter will accept a given value with a known value of the measurable parameter. In this formulation of the problem, the structural and mechanical reorganization of the material of the part during processing can be described by a Markov random walk between absorbing levels, which corresponds to the acquisition of the required properties by the part and their irreparable distortion. For a direct assessment of the hidden characteristics of the process by its known manifestations, a hidden Markov model of the considered structural and mechanical reorganization is proposed, using a continuous phase space of states. On its basis, formulas were obtained for the probabilities of favorable and unfavorable outcomes of processing, which make it possible to introduce a control action into the process at its early stages in order to increase the proportion of the yield of suitable products.