To estimate the impact of fish farming operated at household level on nutritional status among children 6-59 months of age, a cross-sectional study was conducted in Zomba district, Malawi. Anthropometric measurements of 66 children in each type of household (fish-farming and non-fish-farming households) and structured interviews with their parents were undertaken. A total of 21 background variables were employed and examined using bivariate and multivariate analyses. Overall, a lower prevalence of malnutrition was detected among the children in fish-farming households than those in non-fish-farming households in all the malnutrition indicators, i.e. stunting, underweight and wasting. In particular, a significantly lower prevalence was detected among the children in fish-farming households than those in non-fish-farming households in both severe (P=0.045) and global underweight (P=0.042). 'Higher proportion of income from fish farming to total income', 'more frequent intake of oil and fats other than never/seldom' and 'breastfeeding practice for the appropriate duration' are the protective factors against being underweight. Household fish farming may have indirectly contributed to lower prevalence of underweight through increasing frequency of intake of oil and fats by strengthening households' purchasing power. The study supports 12 months as the threshold for appropriate breastfeeding duration.