We present an entropy-based method to quantify a decision-maker's (DM's) subjective utility from an attribute value. Based on such utility values, we develop logit models of discrete choice, termed as entropic logit (EL) model. A critical review of the popular and conventional logit model is presented. In the light of the same, the proposed model is compared with the logit model to highlight the value additions brought by the proposed EL model. EL model is futher extended as compensative EL (CEL) to take into consideration a DM's unique attitude. The proposed EL and CEL models are applied to compute the choice probabilities in a real case-study. IEEE