In forging operations, preform design is one of the most important stages of forging. The shape of the primary raw material for most forging operations is in simple forms such as a cylinder or a cube. If the final shape of the workpiece is complex, the raw material can not be transformed into a final piece at one stage of deformation. In order to prevent problems such as inappropriate material flow, large forces applied to the die and its improper filling, the work piece is deformed into a preform before being formed in the final shape in several stages. The most important criterion for the design of the forging process is that that the preform should fill the final die. However, the shape of the preform still may not be optimal. Contact stresses on the tools and die life need also to be optimized. In this thesis, various stages of forging process of a construction scaffold section have been investigated and affecting parameters that should be considered in the design of forging process are optimized. Analytical simulation has been used as a powerful tool in various design departments. The final model of the workpiece is modeled using the Catia software. Having this information, the final cavity is designed. Simulations reveals that forging of the workpiece in one stage is not feasible due to high stresses in the die. Hence it was concluded that a preform was necessary. Then a preliminary geometry was proposed for the preform, and a test design procedure by response surface was used to obtain preform shape. The simulations were performed by Deform-3D software. Various affecting parameters, such as die cavity filling, thickness of flash, contact stresses and forging force were investigated. Then, using the extracted data, neural network was trained and optimized. Using the genetic algorithm, the output functions were optimized and the proper dimensions were extracted. Keywords: Forging, Hot forging, nuaral networks, genetic algorithm, Deform 3D, simulation