This study examined the criteria used by prospective mathematics teachers in determining whether a problem was a mathematical modeling problem. Two data collection tools were used in this case study. The first was a written evaluation form which consists of four problems developed by 27 prospective teachers individually that enrolled in a Mathematical Modeling course. Among them, only one was a mathematical modeling problem and the others were traditional problems with some features of mathematical modeling problems. The other data collection tool was a semi-structured interview form which was used in the interviews with 12 prospective teachers. The data were analyzed using open and axial coding. The findings showed that pre-service teachers had some incorrect or incomplete information about mathematical modeling. The criteria used by teacher candidates were grouped under three categories: content characteristics, formal characteristics and outcome-based characteristics of the problems. The most common criteria used by the prospective teachers in evaluating the problems was the suitability of real life, being complex or thought-provoking and being interesting. In addition, it was determined that teacher candidates did not adopt standard criteria and that some of the criteria were shaped according to the problems.