Health-related fitness values during childhood (from age 11 to 16 years) were analyzed in relation to changes in physical activity level. One hundred fifty-eight children were monitored over a 4-year period. Twice a year, they performed six physical fitness tests: standing broad jump, 10 × 5-meter shuttle run, sit-and-reach, handgrip, number of sit-ups in 30 s, and 20-m shuttle run. Height was measured four times per year to assess age at peak height velocity. Physical activity was assessed with a self-administered questionnaire at baseline and at the end of the follow-up. The physical activity level was defined by the frequency and the time spent in moderate-to-vigorous physical activities. Two groups were constituted: regularly active and sedentary. A multiple linear multilevel regression analysis was used to analyze the longitudinal relationships between changes in physical activity and physical fitness levels. Corrections were made for both time-dependent (time) and time-independent (sex) variables. Except for flexibility, boys' fitness performances increased more than that for the girls. Positive and significant (P < 0.05) regression coefficients were found with the regularly active for standing broad jump, 20-m shuttle run, number of sit-ups, 10 × 5-m shuttle run in both sexes, and for the girls' sit-and-reach performance. Increasing or decreasing physical activity level was not associated with changes in fitness performances over time, except for flexibility for the girls and the 20-m shuttle run for the boys. From childhood to adolescence, increasing physical activity is not sufficient to be more fit. The children who are stayed the most active were the fittest, particularly the girls. © 2005 Wiley-Liss, Inc.