In view of the fact that the rolling shape of the 20-high rolling mill is affected by multiple factors, it is difficult to accurately predict,based on finite element and PSO-BP method, a prediction model of the rolling shape quality of a 20-high rolling mill was established. According to the positional relationship between the rolls of the 20-high mill, based on the finite element software ANSYS/LSDYNA, a finite element model of the roll system of a 20-high rolling mill was established, considering the factors such as roll elastic deformation, strip plastic deformation and friction. The influence of plate width, thickness, tension and speed on the shape index was analyzed. Considering the different influencing factors of rolling shape, the quality prediction model of rolling shape of a 20-high rolling mill was established based on BP neural network, in which the shape index was used as the measurement index of shape quality. The particle swarm optimization algorithm was used to optimize the weights and thresholds of the BP neural network of the shape quality prediction model, then the shape prediction accuracy is improved.