Often in real-world decision making, it is difficult to crisply evaluate the utility values as required in the case of conventional choice models. Besides, a decision maker (DM) has his/her own relative importance for each of the attributes. The attributes may also be interacting positively (synergy) or negatively, the degree of which is specific to the DM. A new family of discrete choice models is introduced with a motivation that takes into account the human factors in real-world multiattribute decision making. More specifically, the proposed choice models are based on fuzzy subjective utilities that are easier to elicit. The proposed models are further extended to take into account the unique attitudinal character of the DM, the relative weight vector, and the degree of interaction among the different attributes. A real case study illustrates the usefulness of the study. © 1993-2012 IEEE.