Purpose: In this study the fitness of a path model was examined for relationships among loneliness, motives for gambling, gambling risk factor, self-efficacy to control gambling, and low level gambling behavior in college students. Methods: Participants were 410 college students selected from 4 universities and data collection was done between September 22 and October 24, 2014 using self-report questionnaires. Data were analyzed using IBM SPSS 19.0 and AMOS 20.0 programs. Results: Motives for gambling, self-efficacy to control gambling, directly affected college students’ low level gambling behavior, while loneliness, gambling risk factor, gambling motives affected it indirectly. This study also showed that the modified path model is efficient and appropriate to analyze college students’ low level gambling behavior. Conclusion: These results suggest that college students’ low level gambling behavior can be decreased by reducing loneliness, gambling motives, gambling risk factor and increasing self-efficacy to control gambling. Thus, in order to prevent college students’ low level gambling behavior, there is a need to design an intervention program which focuses on reducing loneliness, gambling motives, gambling risk factor, and on reinforcing self-efficacy to control gambling.