Talent identification is a process through which those with the highest chance to succeed in the future will be determined. The purpose of the present study was to investigate the application of a multivariate approach in talent identification of adolescent soccer players. The research population consisted of all soccer players in Tehran province aged between 12 and 16 years old. A battery of tests was applied to measure the psychological, technical, physiological and anthropometric factors in adolescent soccer players in two levels of elite (N=26) and sub-elite (N=35). Participants were divided into four age groups: U16, U15, U14, and U13. The OMSAT – 3 was used to evaluate the mental skills. Furthermore, technical skills were measured by six soccer skills tests of Football Association of England. Physiological measurements consisted of vertical jump, sit up, 10m sprint, 40m sprint and 280m shuttle run. Anthropometric characteristics were height, weight, body mass index, length of lower and upper leg, skin folds in four regions and body fat percentage. Statistical method to analyze the findings was MANOVA. The results showed that among four groups of factors, the physiological measures were statistically more capable of distinguishing the levels. The results of the present study showed a significant difference between elite and sub-elite soccer players in physiological, technical and psychological characteristics, but there was no significant difference in anthropometric attributes. There was a significant difference between elite and sub-elite soccer players in sit up, shuttle run, 10m sprint, 40m sprint, running with ball, turning, speed, dribbling, shooting, goal setting, relaxation, coping with stress, activation, focus, refocus and competition plan (P<0.05). These findings demonstrated that there is more need to emphasize the importance of mental skills in talent identification models. In conclusion, a multivariate approach with regard to the age differences can successfully discriminate elite players from their sub-elite counterparts in adolescence period. It can be suggested that in talent identification process in ages under 13 and 14, physiological measurements are less used and technical characteristics are more emphasized.