In this paper, it is aimed to estimate the daily demand for meal in the places where mass food is produced and thus to avoid overproduction and to prevent the end of meals with less food production. For this purpose, two different regression models were designed as gauss process regression (GPR) and regression trees and the success of the models were compared. In the study, the number of people eating was estimated using data obtained from the university dining hall. For the developed GPR prediction model rational quadratic GPR, square exponential GPR, matern 5/2 GPR and exponential GPR methods were used. In the regression trees prediction model fine tree, medium tree and coarse tree methods were used. The best results were obtained from the exponential GPR and fine tree models.