<i>Objective: </i>To describe the effects of URMEL-ICE, a German school-based intervention for overweight prevention, on children’s BMI and other measures of fat mass. <i>Methods: </i>A cluster-randomised controlled design was used. The intervention which focused on physical activity, TV time and soft drink consumption was integrated into a second-grade curriculum and was implemented by classroom teachers themselves. It comprised 29 teaching lessons, 2 short exercise blocks per day and 6 family homework lessons. BMI was assessed as primary outcome measure, waist circumference and skinfold thickness as secondary outcomes. Data of 945 children were analysed. <i>Results: </i>Multivariate analyses adjusted for baseline values showed no statistically significant effect of the intervention on BMI, but on waist circumference (–0.85; 95% confidence interval (95% CI) –1.59 to –0.12) and subscapular skinfold thickness (–0.64; 95% CI –1.25 to –0.02). After additional adjustment for individual time lag between baseline and follow-up, these effects were reduced to –0.60 (95% CI –1.25 to 0.05) and –0.61 (95% CI –1.26 to 0.04) and lost their statistical significance. <i>Conclusion: </i>This study contributes to the field of randomised school-based studies on overweight prevention and shows that within a 1-year, integrated intervention no effect on BMI, but a tendency towards effects on fat mass can be achieved.