The shear wave velocity of soil is an important physical quantity in soil dynamics, which is indispensable for the determination of site classification and the site soil layer thickness. In recent years, wide application has been achieved in engineering survey, geological engineering, and earthquake engineering by testing the shear wave velocity of soil layers. Previous studies have shown that the value of shear wave velocity changes within a very small range and has certain statistical rules in certain areas with the same sedimentary environment and regional geological background. Moreover, using the statistical relationship between soil depth and the parameters of soil character for exploration is allowed by relevant specification and can thus be viewed on a theoretical and practical basis. By using various statistical methods for fitting, scholars worldwide have determined that the regular empirical relationship of shear wave velocity changes with depth in some areas. Xi'an is one of the oldest cities in the world and has a vivid and rich history and culture. During its 3,100 years of development, it served as the capital for 13 dynasties such as Qin, Han, and Tang. The three main geomorphic units of Xi'an city include the alluvial plain of the Weihe River (area I), the alluvial plain of the Chan-Ba River (area II), and the proluvial-lacustrine platform before the loess tableland (area III). In this paper, the measured data of shear wave velocity of soils from 309 boreholes obtained during seismic safety evaluation of the engineering site in Xi'an city are used to qualitatively discuss the empirical relationship between soil shear wave velocity and soil depth in Xi'an. Shear velocity variation that changes with depth is revealed by using a linear fitting model, a quadratic polynomial fitting model, and an exponential fitting model. The results show that the exponential and linear models can produce effective results within a depth of 20 m and depths of more than 20 m, respectively. Thus, a reliable method includes a combination of both. We adopt the exponential model to fit the former and the linear model to fit the latter regions. Moreover, we analyze the relationships among the shear wave velocities of various sedimentary types, including fine sand, medium sand, coarse sand, loess, silty clay, ancient soil, and pebble soil, with the soil layer depth by using the statistical regression method. Subsequently, we obtain the empirical formula of the soil shear wave velocity (VS) with soil depth (H) in different geomorphic units. Our results show that the value of correlation higher than 0.9 accounts for 29.4%, which indicates a very close relationship between VS and H; that between 0.8 to 0.9 accounts for 55.9%, which indicates a close relationship between VS and H; and that between 0.7 to 0.8 accounts for 14.7%, which indicates that VS is associated with H. The best correlation value was 0.9 475 (function correlation coefficient is 1), and the worst correlation coefficient value was 0.704 5 (0.7 is the threshold value of correlation). Only five of the correlation coefficients were less than 0.8, which indicates that the statistical relationship is relative, and good correlation accounts for the majority. A comparison of the measured and predicted values indicates that the discrepancy is very slight. The results show that the empirical formula we have determined between the VS of various types of soil layers in different geomorphic units and the H is reliable and therefore fits the characteristics of localized soil and rock layers. Therefore, our formula can be used for general construction engineering survey and seismic safety evaluation in Xi'an city.