The hardening strain model with small strain (HSS) can reflect the high nonlinearity of soil modulus at small strain range, which has been widely applied in deformation calculations during pit excavations under complex geological conditions. The appropriate values of model parameters have a significant impact on the computational results. However, the current recommended parameter values have a large range of fluctuations, which brings inconvenience to engineering applications. Therefore, further research is needed on the parameter values of the model. This study analyzes the data characteristics of existing HSS model parameters and introduces the artificial neural network method (ANN). The empirical relationships between the modulus parameters of the HSS model for Shanghai soils and factors such as the initial pore ratio (e) and oedometer modulus (E_s1-2) are established. This allows for the rapid determination of HSS model modulus parameters based on survey reports. Compared with the recommended values, the proposed method for determining the HSS model modulus parameters is closer to the measured value. The study results can provide references for projects in Shanghai.