To deal with the running deviation of thin strip in the continuous annealing production process, a multi-objective operation optimization model with the consideration of safety constraints is established, and a multi-objective evolutionary algorithm is proposed, in which the classification and the multi-population with competition and coordination are incorporated. The mechanism of multi-population with different evolution strategies is established to enhance the search diversity, and the competition and cooperation based on information sharing is adopted to improve the robustness of the algorithm. The solutions in the external archive(EXA) are classified into different groups, and then local search is performed within each group, which can help to ensure the diversity of the non-dominated solutions in EXA and improve the convergence speed of the algorithm. Computational results based on benchmark problems show that the proposed algorithm has good convergence and diversity, and further results based on the practical operation optimization of continuous annealing illustrate the effectiveness of the proposed algorithm.