An evaporation pond is an environmentally friendly method of treating brine. Accurately predicting the evaporation rate of the brine is important for estimating the required site area and the treatment capacity of the evaporation pond. This study presents three empirical models (i.e., one polynomial regression and two machine learning models) and an analytical model (Penman model), for predicting the rate of brine evaporation from different meteorological factors. To develop the models, the brine evaporation rate was measured under 51 different conditions that were carefully controlled in a lab-scale chamber designed to vary air temperature, relative humidity, wind velocity and sunlight illuminance. The three developed models showed excellent internal validity. In addition, the Penman model was modified by correcting two input variables (i.e., wind velocity and sunlight illuminance) to better represent the experimental data. To verify the external validity of the models, outdoor evaporation experiments were performed. The prediction values of all four models were generally well in line with outdoor measurements of both the evaporation rate and its cumulative evaporation weight. The Penman model exhibited the highest accuracy in predicting time-dependent variations of the evaporation rate, while the polynomial and artificial neural network models better explained the cumulative evaporation weight.