Stirring tanks are regarded as pivotal equipment for solid-liquid suspension operations in industrial processes such as adsorption and crystallisation, and its performance is complex in relation to the pulp phase, structure and process conditions. However, the traditional design method for stirring tank structure is sophisticated and time-consuming, which greatly limits the design efficiency. To address the high cost and inefficiency of traditional design methods, this paper performed a multi-objective optimal design of stirring tank based on numerical simulation, surrogate model, NSGA-II, Entropy Weighting Method and Analytic Hierarchical Process. Firstly, the flow field state of stirring tanks with different structural dimensions and operating parameters was simulated with computational fluid dynamics (CFD), which was subsequently validated by Particle Image Velocimetry. Then, for the key parameters including stirring speed, stirrer diameter, stirrer height from the bottom and paddle pitch, experiments were designed using the Latin Hypercube method to establish a Kriging model based on the gained simulation data. Afterwards, the Pareto front was obtained from the optimization using NSGA-II with two objectives of stirring power and suspension uniformity. The result of Analytic Hierarchy Process and Entropy Weighting Method showed that the selected optimal solution can improve the suspension uniformity by 0.7 % while reducing power consumption by 48.3 % compared to the initial design.