The aim of this study was to examine the stability of objectively measured physical activity in Swedish children and to study variables that predicted physical activity and body mass index standard deviation score (BMI SDS) at follow-up. A total of 97 children provided valid repeated measurements of physical activity between 2002 and 2005. The children were on average 7.5 years at baseline (SD+/-0.92) and 9 years at follow-up (SD+/-0.92). The mean follow-up time was approximately 1.5 years (mean 558 days, SD+/-224). An accelerometer (Actiwatch, Cambridge Neurotechnology Ltd., Cambridge, UK) was used to measure physical activity during 7 consecutive days. Yearly weight and height were examined and BMI SDS was calculated. Baseline physical activity was significantly correlated with physical activity at follow-up (r=0.59) with a stronger correlation for boys (r=0.72) than for girls (r=0.51). High physical activity levels were more stable (r=0.74) than low physical activity levels (r=0.55). Physical activity at follow-up was explained by physical activity at baseline and season (R(2)=0.46) whereas BMI SDS at follow-up was explained by BMI SDS at baseline and age (R(2)=0.90). The results of this study suggest that physical activity levels are fairly stable in 6-10-year-old children.