Abstarct: Earthquakes are one of the most destructive natural disasters that have ever threatened human and financial communities. Iran is one of the earthquake-prone countries in the world, so that every year it experiences many small to medium earthquakes and every ten years it experiences a major earthquake. Due to the earthquake situation and the special geographical and climatic situation, Iran has the potential to cause numerous and severe earthquakes. Therefore, seismic studies and construction conditions are necessary. An overall assessment of the damage to structures and facilities in earthquakes shows that the construction conditions have had a significant impact on the distribution of damage in cities and villages. Therefore, it is necessary to identify the effect of the construction site, the type of soil and its impact on earthquake waves in the design and construction of resistant structures. Determining the type of soil in a seismic record area is a laborious and costly task. The aim of this study was to predict the relationship between determining the seismic shear wave velocity using the gene exxpression program (GEP) and determining the soil type of the region using the horizontal to vertical (H/V) spectral ratio method. Factors influencing the magnitude of the earthquake include the distance from the source of the earthquake to the structure, the spectral ratio of the horizontal to vertical component (H/V), and the time period of the maximum spectral ratio. In this study, large torque was used as the magnitude and focal length as the distance between the earthquake source and the construction site. The first step in the research is to collect acceleration data, which used 480 accelerometers, and the required corrections were made to these data, and then the input parameters mentioned were extracted to determine the correlation of the shear wave velocity. After preparing the earthquake catalog, the smart method of the gene development algorithm was finally used to obtain the predictive relationship. The advantage of this algorithm is that the fixed regression model is not used and the model is calculated intelligently. The computational results have a fitness of 911.57 and correspond to 73% with the predicted values, which provides a good knowledge of the soil condition of the region compared to the exact method, ie laboratory work, faster, more affordable and at a much lower cost.