Objectives The aim was to examine the changes in metabolic syndrome risk factors over a 2‐year period, and to investigate the independent influence of baseline physical activity (PA) and cardiovascular fitness (CVF) on these changes. Methods 120 Estonian boys (age at baseline 11.9 ± 0.1 years) were grouped according to baseline PA or CVF/kg (VO2max/kg) and CVF/LBM (VO2max/LBM). PA was assessed by 7‐day accelerometry. Total cholesterol (TC), high‐density lipoprotein (HDL), triglycerides (TRG), insulin and glucose were measured and assessment for insulin resistance (HOMA‐IR) and TC/HDL ratio were calculated. Results In both CVF/kg and CVF/LBM, the low CVF groups had significantly higher values of HOMA‐IR (P < 0.009) over time. In TRG and TC/HDL ratio values the only significant difference over time emerged between CVF/kg groups (P < 0.001). Participants in high metabolic risk CVF/kg group were 5.9 times more likely to have high HOMA‐IR values, 2.9 times more likely to have high triglyceride values, and 3.5 times more likely to have high TC/HDL ratio values (P ≤ 0.045) in the second year follow‐up compared to those who were in the low metabolic risk CVF/kg group. In moderate‐to‐vigorous PA groups there were no significant differences between HOMA‐IR, TRG, and TC/HDL ratio values over time. Conclusions The results of the study indicate that CVF has a stronger longitudinal prediction value compared to moderate to vigorous physical activity in terms of metabolic risk factors in adolescent boys. Fitness remained a significant predictor if the influence of body fatness was removed from the analysis. Am. J. Hum. Biol. 28:782–788, 2016. © 2016Wiley Periodicals, Inc.