The purpose of this study was to (1) create a predictive model that would correctly identify men at greatest risk of dropping out of a court-mandated, batterer intervention program; and, (2) explore the creation of such a logistic regression model using a set of instruments that were different from those used in previous research. Method: The study employed a secondary analysis of 114 men, 56 treatment completers and 58 drop-outs, randomly selected from a larger pool of 250 men court ordered into treatment for domestic violence offenses. Results: Analysis indicated that very few demographic and psychological variables differentiated between treatment completers and drop-outs. However, a logistic regression model was developed that correctly predicted treatment completion for 60% of the sample. Conclusion: Implications of the findings for improving retention rates among men attending court-mandated batterer treatment programs were explored and discussed. (Contains 4 tables.)