Considering high efficient searching ability of PSO and local searching capability of SA, a hybrid PSO algorithm for solving the problem of machining processes balancing was proposed. Aiming at the characteristics of operations order and clamping constraints, an improved way of encoding and decoding was presented, which makes the coding sequence of discrete matches with the iterative evolution of continuous particle position and speed, at the same time, a method of dealing with such operational constraints was also stated, which makes the updated particles always satisfy the constraints, the convergence rate of the algorithm was improved greatly. Finally, a case of diesel engine cylinder block machining production line was illustrated to prove the validity of the proposed method.