To investigate the relationships between physical activity, dietary intake and body composition in children. A cross-sectional study on physical activity, nutritional intakes and body composition conducted in 86 healthy 10 y old French children. In addition, growth parameters and nutritional intakes were available from the age of 10 months. Physical activity level (using a validated activity questionnaire over the past year), nutritional intake (dietary history method), anthropometric measurements (body weight, height, arm circumference, triceps and subscapular skinfolds, Body Mass Index (BMI), arm muscle and arm fat areas calculated from these measurements) at the age of 10 y. Anthropometric measurements and nutritional intakes were recorded in the same children at the age of 10 months and every 2 y from the age of 2 y. At the age of 10 y, active children ingested significantly more energy than less active children, mostly due to higher energy intake at breakfast and in the afternoon. This higher energy intake was accounted for by increased consumption of carbohydrates (281 g vs 246 g; 49.6% vs 47.4% of total energy). Even if the amounts of fat consumed were similar in both groups (90 g vs 84 g; P = 0.09), the percentage of fat intake was lower in active children (35.4% vs 37.4%; P = 0.04). The percentage of protein was not different (14.9% vs 15.3%; P = 0.33). In spite of a higher energy intake in the active group, active and less active children had similar BMI at the age of 10 y. However, their body composition differed significantly: active children had a higher proportion of fat-free mass, a lower proportion of fat-mass as measured in the arm and they had a later adiposity rebound. Fatness was significantly and positively associated with the time spent watching television and video games. Physical activity was associated with improved body composition and growth pattern. This association may be related to nutritional changes: active children consumed more energy by increasing carbohydrate, thus reducing the relative fat content of their diet. These results provide support to encourage physical activity during childhood.