Abstract: Unidirectional tensile creep tests of A356 aluminum alloy were carried out at temperatures of 150, 250, 300 ℃ and stresses of 20-150 MPa to investigate its creep characteristics. The multi-objective particle swarm algorithm was used to obtain the parameters of the Norton-Bailey power law model, Garofalo hyperbolic sinusoidal model, and deformation mechanism-based true-stress creep model, and the minimum creep rate-stress curve of A356 aluminum alloy was predicted by these three models. The prediction accuracy of the three models was analyzed. The results show that the minimum creep rate of A356 aluminum alloy was positively correlated with the test temperature and stress, and the temperature had greater influence on the minimum creep rate than the stress. The deformation mechanism-based true-stress creep model had the highest prediction accuracy, and 90% of the data points in the prediction results of the minimum creep rates were located within the 2 time error band. The deformation mechanism-based true-stress creep model obtained by the multi-objective particle swarm algorithm could accurately simulate the creep properties of A356 aluminum alloy under a wide range of temperature and stress.