Abstract: Because structural parameter matching has a strong influence on the service behavior of planetary roller screw mechanisms (PRSM), understanding how to effectively design the structural parameters of PRSM is highly important in practical industrial applications. This study proposes a parameter optimization model based on a crow search algorithm (CSA) to solve the structural parameter matching problem of PRSM. The relationship between the main structural parameters of PRSM can be deduced according to the working principle and geometric conditions. The screw, roller, and nut space spiral surface equation are established by considering the influence of the meshing point position of PRSM on thread meshing clearance. The relationship between the meshing point position and PRSM tooth thickness is obtained using the tangent contact condition of the spiral surface. To achieve no backlash meshing and improve PRSM transmission accuracy, the thread tooth thickness of the screw, roller, and nut can be adjusted. According to the meshing relationship between the thread pair and the gear pair of PRSM, the structural parameters of the annular gear and the gear at the end of the roller are determined. The normal vectors at the meshing point of the screw, roller, and nut are calculated using space spiral surface equations. To test the strength of the PRSM, static analysis of the roller is used to deduce the force relations between the main parts of the PRSM. PRSM structural parameters serve as design variables. An optimization goal is to reduce the outer diameter of the nut, the nominal diameter of the screw, and the length of the roller. The spatial structure constraints and component strength constraints of PRSM are considered. The CSA was introduced to be the optimization algorithm. The parameter optimization model of PRSM is established for achieving optimum matching of the optimization variables. Finally, using the proposed optimization model, three groups of PRSM structural parameters are obtained for three different types of load. In this study, the optimization results are compared with a foreign PRSM product manual to validate the effectiveness of the optimization model. The results show that the PRSM structural parameters obtained from the proposed model are essentially consistent with those from the foreign PRSM product manual. Furthermore, the proposed model provides the structural parameters of thread teeth, which are not included in the foreign product manual. The proposed PRSM optimization model is promising for its application in actual production.