The aim of this study was to examine relationships between results of soccer-specific skill tests and game-related soccer skills in young players aged 12 and 15 years. The participants were 60 male soccer players aged 11.87-14.76 years divided into two groups: lower secondary school candidates (12y) aged 11.87 ± 0.23years (n = 30), and upper secondary school candidates (15y) aged14.76 ± 0.42 years (n = 30). Soccer-specific skills were assessed using two outdoor tests in competitive match conditions: Dribble Test and Turning Test. The assessment of the game-related skills was carried out by five expert judges during a 5 vs 5 small-sided game. The evaluation criteria had been set by a panel of experts. The reliability of game-related skills assessments procedures was calculated using intra class correlation (ICC), standard error of the mean (SEM), coefficient of variation (CV). Spearman’s rank correlation coefficients were used to determine the relationships between results of assessments of game-related skill and soccer specific skill tests. Relationships between soccer-specific skills and game-related skills in 15-year-old players (–0.325 to –0.452) are less pronounced than in 12-year-old players (–0.496 to–0.667). It seems to be unjustified to use soccer-specific skill tests for prediction of young soccer players’ potential. The results of these tests can depend on numerous biological differences in the development of adolescent players. In conclusion, evaluation of players by judges should be part in talent identification models in youth soccer