<p>A sudden increase in the radial depth (SIRD) is a distinctive phenomenon in plunge milling. It is typically characterized by a sharp increase in cutting force at the end of the axial feed of the tool, accompanied by harsh machine vibration sounds, which can negatively impact the reliability of plunge milling. This paper proposes an optimization method to eliminate SIRD in five-axis plunge milling. Initially, a five-axis plunge milling experiment and an analysis of the spatial position relationship between the plunge tools and the workpiece revealed that the cause of SIRD is unreasonable tool path planning. Subsequently, using the cutter position and cutter axis vector as variables, an SIRD discrimination model was developed for adjacent cutter positions and extended to multiple cutter positions. Optimizing the plunge milling tool path is considered a multivariate optimization problem that involves determining the cutter point and cutter axis vector. The SIRD discrimination model was used as a constraint function to aid in solving for the variables. The simulation and experimental results indicate that with the remaining volume of material as the optimization target, the optimized plunge milling tool path results in a residual material volume that is less than 60% of the gradually decreasing plunge depth. This optimization decreases the subsequent semi-finishing time of the workpiece and enhances machining efficiency. Additionally, it does not rely on operator experience and facilitates efficient automated optimization of the tool path to exclude SIRD.</p>