Introduction. Importance of physical activity in psychophysical health preservation is well established, and increasing physical activity level of insufficiently active student population should be goal of public health actions. For them to be efficient, it is necessary to determine all factors, including psychological, that are related with physical activity level. Combination of psychological factors which, in prior research, proved to be succesful in predicting different kind of behavior, including physical activity, and which are based on theories of planned behavior, self-determination, motivation of achievement and HEXACO personality theory, should result in optimal predictive solution. The aim of this study was to determine relationship between leisure-time physical activity level and psychological factors (motivation variables and personality traits) by creating predictive structural model of student physical activity. Methods. Participants were 1304 students from ten University of Zagreb colleges that covers five different scientific areas. There were 857 female students (65,7%) and 447 male students (34,3%). Measure instruments used in condacted research were as follows: Croatian version of International Physical Activity Questionnaire (IPAQ, long form), Theory of planned behavior questionnaire regarding physical activity, Croatian version of Task and Ego orientation in sport questionnaire, Croatian version of EMI-2 questionnaire, the Locus of Causality for Exercise Scale, and Croatian version of HEXACO-PI-(R) personality measure. Results. Leisure- time physical activity can directly be predicted by intention to be physically active, by locus of causality for exercise and by group of social motives for exercise, with higher intentions, inner locus of causality and greater motivation by social motives leading to higher level of student leisure-time physical activity. Furthermore, student leisure-time physical activity level prediction by personality traits and motivation variables depends on student physical activity level. Final model solution created from sample of students who fullfil minimum leisure-time activity level recommendation explained 13,3% of student leisure-time physical activity variance, while final model solution created from sample of students who don't fullfil minimum leisure-time activity level recommendation explained 3,9% of student leisure-time physical activity variance. Results showed that, considering physical activity level, there is statistically significant difference in attitudes toward physical activity, subjective norms of physical activity participation, perceived control of physical activity, intention to be physically active, task goal orientation, locus of causality for exercise, agreeableness, exstraversion, emotionality, psychological motives for exercise and social motives for exercise. Considering gender, results showed that there is no difference in successfulness of predictive models, with final model solution created from male students sample explaining 16,5% of student leisure-time physical activity variance, while final model solution created from female students sample explaining 16,3% of student leisure-time physical activity variance. But, considering gender, there is statistically significant difference in attitudes toward physical activity, perceived control of physical activity, task goal orientation, ego goal orientation, locus of causality for exercise, openness to experience, conscientiousness, agreeableness, emotionality, honesty-humility, health motives for exercise, social motives for exercise and, in the end, in leisure-time physical activity level and total physical ctivity level. Conclusion. Theory of planned behavior variables proved to be most consistent leisure-time physical activity predictor, with exception of perceived control, which, contrary to theory assumptions, predicted only intention of physical activity, not behavior itself. Other variables that proved successful in predicting students leisure-time physical activity, regardless of analysed sample, were locus of causlity for exercise, social motives for exercise, psychological motives for exercise, ego goal orientation and agreeableness personality trait. Relatively small percentage of explained student leisure-time physical activity variance by psychological factors indicates a need to include factors from other fields (sociodemographic, sociological) in predictive model with wich we would be able to detect those most efficient, so most successful interventions for increasing physical activity level could be created.