OBJECTIVE: To test a surveillance system on diet and physical activity based on data gathered at well-child visits. DESIGN: Cross-sectional data collection on growth, diet, physical activity and sociodemographic variables. SETTING: Offices of 179 paediatricians in three Italian regions. SUBJECTS: 26 898 children for a total of 32 915 well-child visits at 1, 3, 5-6, 8-9, 12, 18, 24, 36 and 60-72 months of age. RESULTS: The BMI Z-score was lower than the WHO standard at 1 and 3 months but higher from 8-9 months onwards. The rates of breast-feeding at 1, 3, 5-6, 8-9 and 12 months were 88 %, 75 %, 64 %, 52 % and 32 %, respectively, with 5 % and 3 % continuing up to 24 and 36 months. Exclusive breast-feeding was 64 %, 54 % and 20 % at 1, 3 and 5-6 months, respectively; 57 % at 5 months and 85 % at 6 months were given complementary foods. Only 8 % and 10 % of children were taking five portions of fruit and vegetables daily, while 47 % and 51 % were consuming sugar-sweetened beverages at 36 and 60-72 months, respectively. At 60-72 months, less than 10 % reported at least 1 h of moderate-to-vigorous physical activity on 5-7 d/week, and 32 % watched television or played videogames for more than 2 h/d, every day. The majority of paediatricians rated the surveillance system as reliable and feasible. CONCLUSIONS: Surveillance for diet and physical activity in pre-school children, with data gathered during well-child visits, is feasible and potentially useful to plan and evaluate activities for the prevention of obesity.