This research provides an alternative framework for the analysis of employee stock option exercise patterns. It develops a binomial model where the exercise decision obeys to a policy that maximizes the expected utility to a representative employee exhibiting preferences as described by the Cumulative Prospect Theory (CPT). Using a large database on exercise transactions in 12 US public corporations, I examined the performance of the model in predicting actual exercise patterns. Interestingly, the probability weighting coefficients yielded by the model calibration are consistent with those from the experimental literature. Further, the results suggest that the model outperforms the Expected Utility Theory-based model in predicting actual exercise decisions in the sample. These findings convey the main contribution of this paper: the strong ability of the CPT framework to explain employees exercise behavior. It therefore provides rationale for using this framework in order to get more relevant fair value estimates of stock options.