To address the issues of low mixing uniformity and high drilling resistance in traditional mixing pile drilling tools, a novel optimization design method based on an improved multi-objective particle swarm optimization algorithm is proposed. Firstly, a coupled dynamics model of drilling tools, cement slurry, and soil particles is simulated using a multibody-fluid-particle integrated approach, analyzing the influence of drilling tool structure on pile formation quality. Secondly, a Dual Adaptive Multi-Objective Particle Swarm Optimization (DAMOPSO) algorithm is developed by incorporating adaptive Cauchy mutation and quasi-oppositional learning strategy, combined with an improved adaptive inertia weight and learning factor, thereby enhancing the solution efficiency and accuracy of the optimization model. Finally, taking the number of cement slurry nozzles, the angle and length of cutting blades and the rotation speed of mixing shaft as design variables, and the cement-soil mixing uniformity as well as the drilling resistance as optimization objectives, a multi-objective optimization model for the mixing pile drilling tool is established. By integrating the DAMOPSO algorithm with grey relational analysis, the optimal design of the mixing pile drilling tool is obtained.