We conducted a systematic review and meta-analysis to compare the effect of breast-feeding and formula-feeding on body composition of preterm infants. We searched the literature using PubMed, Cochrane Central Library Issue, Ovid (Medline), Embase and other resources such as Google Scholar, electronic databases and bibliographies of relevant articles; two reviewers collected and extracted data independently. All the authors assessed risk of bias independently using the Newcastle–Ottawa Scale (NOS). A fixed-effects meta-analysis was undertaken with RevMan 5 software (The Cochrane Collaboration) using the inverse variance method (P≥0·05; χ2 test). In contrast, a random-effects meta-analysis was carried out. Altogether, 630 articles were identified using search strategy, and the references within retrieved articles were also assessed. A total of six studies were included in this systematic review. In formula-fed infants, fat mass was higher at term (mean difference 0·24 (95 % CI 0·17, 0·31) kg), fat-free mass was higher at 36 weeks of gestational (mean difference 0·12 (95 % CI 0·04, 0·21) kg) and the percentage of fat mass was higher at 36 weeks of gestation (mean difference 3·70 (95 % CI 1·81, 5·59) kg) compared with breast-fed infants. Compared with breast-feeding, formula-feeding is associated with altered body composition from birth to term in preterm infants. The effects of formula-feeding on preterm infant body composition from term to 12-month corrected age are inconclusive in our study. Well-designed studies are required in the future to explore the effects of formula-feeding compared with breast-feeding.