The identification of the possible relation between the factors of retention and academic dropout factors, according to the exploration of individual issues, academic, institutional and economic development in two population samples, with the predominant factors that characterize resilient will be the base for the construction of intelligent agent who favors the risk assessment of desertion. The activity was developed with two population samples, one of 113 active students from cead Jose Acevedo y Gomez from the National Open & Distance University, who applied the questionnaire designed, validated and reliability analysis for retention and CRE-U questionnaire, and another with 53 people who dropped out of the academic process, who applied the questionnaire designed, validated and reliability analysis for desorption and CRE-U questionnaire. The methodology used is quantitative, of descriptive cut, causal comparative study of survey, comparative and correlation. For the survey of desertion factors the results were grouped by factor and the analysis indicates that in all the factors significant differences exist also the results of the resilient factors survey, but except independent factor, they have different behavior at statistical significant differences. The results of this study allow to define a model where, from factors relation, indicate whether an active student is at risk of dropping out.