With the development of intelligent manufacturing, the heterogeneous parallel workshop scheduling problem has increasingly attracted attention. In such workshops, machines vary in number and processing capabilities, requiring coordinated optimization to enhance production efficiency. It addresses the optimization problem of scheduling in heterogeneous parallel workshops by establishing a mathematical model with the objectives of minimizing total production costs and minimizing total tardiness. An improved multi-objective particle swarm optimization (MOPSO) algorithm is proposed to solve this problem. The algorithm dynamically adjusts the inertia weight to enhance the global search capability of the particle swarm in the solution space. It also improves algorithm efficiency by ranking and calculating cycle distances to identify global learning samples. Experimental results demonstrate that the proposed algorithm effectively balances the two objectives and generates Pareto-optimal solutions, providing decision-makers with various options.