This research is concerned with the effects of the geometrical parameters of the die in elevated temperature Hydro-Mechanical Deep Drawing (HMDD) process of 2024 aluminum alloy. A Group Method of Data Handling (GMDH) process was used to train a neural network in order to study the process behavior. Based on the maximum reduction in sheet thickness and the uniformity of the final product, an objective function was constructed. The Bees Algorithm (BA) was used to achieve the optimal values for process variables. To verify the simulation results, they were compared with the experimental findings gained via this research and an appropriate correlation was observed between these results. This comparison showed that, by optimization of the geometrical parameters of the process, the value of the combined objective function was the best one compared with all of the cases tried in the present investigation.