This paper presents a statistical procedure for tool replacement in a two-stage metal cutting process where a workpiece is being machined by two tools, and if one of the tools fails during the cut, the workpiece becomes defective and an interruption of cutting due to excessive tool wear or sudden fracture of a cutting edge occurs. A replacement procedure is formulated from a statistical point of view. Numerical results show that the efficient tool replacement in a two-stage cutting process differs from that of a single-stage one.