A learning model of tool wear based on Bayesian statistical methods provides a means for regulating the optimum cutting conditions as periodic sampling data on flank wear become available during production under adaptive control. The sampling process is used to estimate the current parameters of the wear process, and by incorporating this updated information into the machining economics model, an optimal a posteriori program of cutting conditions can be determined to best match the current conditions of the tool, workpiece, and machine. The application of the Bayesian learning model is illustrated for a basic turning operation with minimum cost as the optimizing criterion.