A radar signal sorting algorithm based on convex clustering is proposed to address the issues of existing radar signal sorting algorithms that rely on manually preset number of clusters and are sensitive to initial input values. Convex clustering algorithm is a clustering analysis method based on objective function， which can ensure the global optimal solution of radar signal sorting task by optimizing the convex objective function， and the clustering effect is independent of radar signal input. After standardizing the pulse description word （PDW） output by the interception receiver， construct a convex optimization objective function and solve it using alternating minimization algorithm to obtain the final sorting result. Simulation experiments show that the radar signal sorting algorithm based on convex clustering achieves high sorting accuracy， while maintaining good stability and noise robustness.