For the container packing scenario of sheet paper cargo(SPC) in the paper industry, this paper investigates the SPC packing problem under real-world constraints and constructs a mixed-integer programming model for SPC packing. To solve the SPC packing model, a set cover model is established, and a Benders decomposition-column generation(Benders-CG) algorithm is proposed. Initial columns are generated based on the skyline algorithm, and the Benders decomposition algorithm is employed to solve the CG subproblems. Numerical experiments are conducted using actual SPC data from a domestic paper mill to verify the solvability of the model and the optimization efficiency of the algorithm. Additionally, based on out-of-sample data, sensitivity analysis is performed to explore the impact of parameters. The results demonstrate that the Benders-CG algorithm outperforms traditional CG algorithm and optimization solvers in instances ranging from small-scale to large-scale. When SPC demand or container costs fluctuate, the algorithm converges stably, with reasonable total costs and an optimal demand interval. This research provides an effective approach for SPC loading optimization in the paper industry.