Aiming at the problems of existing wolf pack algorithms in solving constrained multi-objective optimization problems (CMOPs): imperfect constraint handling mechanism leads to insufficient solution feasibility; single search mode of scouts and fierce wolves fails to fully explore the solution space, resulting in insufficient solution diversity, a constrained multi-objective wolf pack algorithm with spiral search and multi-population interaction (CMOWPA-SSI) is proposed. First, a spiral search strategy is designed to optimize the learning models of scouts and fierce wolves, which strengthens global exploration and local exploitation capabilities respectively, thus improving the diversity and convergence of the algorithm. Second, a multi-population interaction mechanism is constructed, which sets up three populations including constraint-priority, constraint-ignorance and constraint-relaxation. The complementary advantages are realized through information interaction among populations, which coordinates the exploration of feasible and infeasible regions, enhances the algorithm"s adaptability in constrained environments and ensures the feasibility of the solution set. Experiments on DTLZ, MW and DAS-CMOP benchmark test suites show that CMOWPA-SSI is significantly superior to five comparison algorithms in convergence accuracy and solution set distribution, and can effectively improve the comprehensive solving performance of CMOPs.