This research presents a mathematical model for the effect of salinity on growth and yield of soybean. Soybean plants are grown in solution culture or hydroponics with different concentration between 0-30 mM NaCl. Exponential equations are used to predict the dry weight and the ratio of pod and flower. The approximated parameters of the equations are found by trust-region method. In addition, a feedforward neural network model is used to predict the yield of soybean (seed weight). The dry weight, the electrical conductivity, the ratio of pod and flower and the season are used as input variables and the yield is used as the target variable. The result shows that the model can be used to predict the yield with relative error less than 4 percents.