The main goal of this study was to determine the best decision tree algorithm in order to determine the effects of the birth type, herd type, main age, pasture type, sex and lamb color variables in the Hemsin lambs. These data, taken during certain periods of the Hemsin lambs, were subjected to simulation. The obtained data were evaluated by Classification and Regression Tree (CART), Boosting, Bagging (Bag) and Random Forest (RF) Decision Tree Algorithms. The best model of lamb birth weight, with the lowest Error Squares Mean (MSE) = 0.469, with the lowest Mean Absolute Error (MAE) = 0,471 with Random Forest and the lowest Symmetric Percentage Mean Absolute Error (SMAPE) = 3,63 The best model according to daily live weight increases of lambs, the lowest MSE (3620.67), MAE (4878.71) and SMAPE(4.80) values ​​formed by Random Forest algorithm and the best model according to the daily live weight gain constructed by Random Forest (RF) algorithm. It was determined that the Bagging algorithm with the lowest MSE (970.09), MAE (1362.65) and SMAPE (3.03) was formed. Theachieved results showed that the best algorithm was Random Forest, follewed by Bagging algorithm.