Dimensional synthesis was the core content in the parallel mechanism design. Therefore, a dimensional synthesis method based on the improved chaotic particle swarm algorithm was proposed for the Ahut-Delta parallel mechanism. Firstly, improved chaotic particle swarm algorithm was proposed. In the algorithm, initialization of population with chaos cube map was experienced. Then inertia weight was adjusted exponentially on the basis of the algorithm iterative state. Simultaneously, early maturity judgment and chaotic disturbance were utilized to obtain the optimal particle. Secondly, the optimal parameters of Ahut-Delta were changed to the dimensional variables. The population mean condition number and the population fluctuation rate of Jacobian were synthesized to a global performance index, and then the global performance index was changed to the fitness function for the improved chaotic particle swarm algorithm under the geometric constraints and the transmission angle constraints of the Ahut-Delta. Thirdly, an optimization simulation on the scale parameters for Ahut-Delta parallel mechanism was conducted by using two optimization algorithms, i.e., basic particle group algorithm and improved chaotic particle swarm algorithm. Through the analysis of the two algorithms results, the optimal particle with the minimal fitness function value was optimized by means of improved chaotic particle swarm algorithm, and the optimal scales were obtained which remarkably improved Ahut-Delta motion performance. Finally, the results of simulation and analysis verified the correctness and effectiveness of the method.