OBJECTIVE: We measured physical activity changes among 10-year-old British children over 12 months and assessed biological and demographic determinants. METHODS: Physical activity was measured with accelerometers (counts per minute) over =3 days at baseline and 1 year later in a prospective study of 844 children (41.6% male; mean ± SD baseline age: 10.2 ± 0.3 years) from 92 schools. Meeting physical activity recommendations was defined as =60 minutes/day at =2000 counts per minute. Biological (height, weight, and fat percentage) and demographic factors (gender, rural/urban home location, and socioeconomic status) were assessed at baseline. Associations between physical activity changes and multiple factors were studied. RESULTS: Physical activity decreased over 1 year (baseline: 665.7 ± 209.8 counts per minute; follow-up: 623.2 ± 179.2 counts per minute; P < .001), with 70.4% of children meeting physical activity recommendations at the baseline evaluation and 65.8% at the follow-up evaluation (P < .001). The decrease occurred mainly on weekends (-47.2 ± 395.8 counts per minute; P = .002), with no significant change on weekdays (8.0 ± 201.6 counts per minute; P = .20). Girls (P < .001), participants with greater body fat percentage (P = .001), and participants of higher socioeconomic status (P = .008) were more likely to exhibit physical activity decreases. CONCLUSIONS: Physical activity decreased over 1 year among children in primary school, predominantly during the weekend. Because these children were relatively active at baseline, prevention of physical activity decreases in childhood, particularly among girls and on weekends, may be a suitable health promotion target.