Abstarct: Abstract The existing limitations of the Equal-Channel Angular Pressing (ECAP) have been led to the development of methods for continuous processes. These limitations include the short length of the workpiece, the discontinuity of the process and the discarding of the beginning and end edges of the workpiece, the limitation of the stroke, and needed to excess parts for removing the material in the channel of the ECAP die. Due to these factors, the ECAP process is not efficient and cost-effective in mass production. In recent years, researchers have paid special attention to methods of Severe Plastic Deformation (SPD) continuously, such as the ECAP-Conform process. In the present study aiming to a comprehensive knowledge in the ECAP-Conform process, the influence of effective parameters on the ECAP-Conform process of AA7075 aluminum and their optimization have been investigated. The aim is to achieve maximum stress, minimum torque, and minimum rod curvature as well as to result a more uniform strain distribution. Influence of the wheel radius, the bending angle, the die channel angle, the outer corner of die channel angle, the friction coefficient between rod and wheel, the friction coefficient between the rod and die and the dimensional ratio of die channel are examined on required torque, Yield strength, the curvature of the output rod and the uniformity of the strain distribution at the cross-section of the rod. For this purpose, 88 experiments were designed baxsed on the Response Surface Method (RSM) in the Minitab software and their simulations (after validation) were performed using Abaqus software. To model the ECAP-Conform process of the aluminum rod, multilxayer perceptron Artificial Neural Network (ANN) has been used. To find the number of laxyers as well as the number of neurons in each laxyer, the Particle Swarm Optimization (PSO) algorithm and the Whale Optimization Algorithm (WOA) have been used. The optimized two-laxyer perceptron neural network with the PSO algorithm was selected as the best neural network with an mean square error rate of 0.0027. baxsed on it, the optimization of the ECAP-Conform process has been implemented by PSO and WOA algorithms. The results show that the parameters of the die channel angles and the friction coefficient between the rod and the die have a significant effect on all responses. Comparing the results of optimization methods and simulations shows that the PSO algorithm has the lowest error with an error of about 15%, and the RSM and the WOA algorithm, both have an error of 17%.