Classification of pig carcasses according to its characteristics is a determining factor in slaughterhouses, since allows the optimization of production and improves the performance in cutting plants and in other subsequent processes. Usual criteria in carcasses classification are the weight and the fat content of the ham, especially in regions such as Spain where a significant proportion of hams are used for curing to produce Jamón Serrano. The objective of this study is to compare different models based on Decision Trees and using intrinsic data of pigs to classify hams in four groups according to their predicted weight. The model presented is based on Bagged Decision Trees and use as input: the weight and the lean meat percentage (LMP) of the carcass, the sex and the breed of the pig and the manual classification of the ham according to the thickness of the subcutaneous fat. The results show a success rate of 81.7%, improving by 4.4% the results obtained with a more straightforward decision tree based only on the weight and the LMP.