<p>Multi-pass compression deformation experiments for a high-strength container steel have been conducted on the DIL805A/D thermal expansion instrument. The true stress- plastic strain curves of experimental steel were plotted. Three typical flow stress models are used to predict the flow stress of the first pass deformation, and Model-1 flow stress model with the highest fitting accuracy is selected as the basic model form. Also, high precision static recrystallization volume fraction model and austenite grain size model have been established. The genetic algorithm is used to optimize the parameters in Model-1 model according to the second pass flow stress data. The relationships between static recrystallization volume fraction, the initial austenite grain size, the dislocation density before deformation, the deformation temperature, the strain rate and the model parameters are established through the Support Vector Machine (SVM) algorithm. The established flow stress model not only has high accuracy but also conforms to physical metallurgical principles under multi-pass steel deformation conditions according to a maximum plastic strain of 0.25. The research results can provide an important theoretical guidance for the load distribution of the rolling mill for the production of high-strength container plate.</p>