In this research work, an optimization of the design of steel structures is developed through Genetic Algorithms (Gas) with the objective of reducing the amount of material and with them the cost, for which the programming of the GAs tool in the Fortran programming language with the SAP2000 Application Programming Interface (API) v.19 using Visual Studio Professional 2015-Intel® Visual Fortran with Parallel Studio XE. The GAs method consists of randomly searching the Objective Function (OF) of the optimal variables with restrictions. The OF defined by the weight of the structure, the variables in relation to the cross sections of the structural elements and the constraints are the design of resistance using the method of Load and Resistance Factor Design (LRFD), restriction of displacements and restriction of the ratio of slenderness considered in the National Building Regulations (RNE), whose GAs search parameters are given by probability of crossing 70% by simple crossing strategy, probability of mutation 20%, elitism of 5% and maximum number of generations according to the profile. The examined metal structure is a cover with five double-body warren arches with 32 m light connected with 24 joists both with welded connections, that is, rigid. The metal structure was evaluated on demand for loads (dead load, live load, snow load, wind load and earthquake loads). The structural design was carried out with the following profiles: circular, tubular, square, angular and Normal Profile with I-shaped section (IPN) of which the most optimal was the tubular profile. The outer diameters of the cross section are: diagonal with a diameter of 1 in, lower flange with a diameter of 2 3/8 in, upper flange with a diameter of 1 1/2 in, inner belt with a diameter of 1 in and belt. outside with a diameter of 1 1/2 in.