Micronutrient malnutrition is a public health concern in many developing countries including Sri Lanka. Rural poor households are more vulnerable to micronutrient malnutrition due to their monotonous rice-based diet, which lacks dietary diversification(1). Despite the potential of home gardens on increased food access and diversity, their contribution to household dietary diversity remains unclear. This study aimed to investigate the impact of home gardens on diet diversity among rural Sri Lankan households. Low-income households with children under five were randomly selected from the Samurdhi beneficiary list, and 450 households having a home garden agreed to be interviewed. We collected information on types of crops and livestock produced over the past 12 months and their utilisation. We also collected the socio-demographic characteristics of the households. We measured household dietary diversity using the Household Dietary Diversity Score (HDDS) based on FAO guidelines. Multiple linear regression was used to identify the predictors of HDDS. Complete data sets were only available for 411 households and were included in the analysis. The HDDS ranged from 3 to 10 with a mean of 6.4 (±1.37 SD) indicating a moderate level of dietary diversity. However, only 20.4% of the households met the adequacy threshold, which is higher than the third quartile(2). Cereals, and fats and oils were the only food groups consumed by all the households. Although many households produced fruits (67.2%) and reared livestock (48.2%), the consumption of these groups were the lowest among the 12 food groups. Predictors of HDDS included monthly household income which had a strong positive relationship, especially earnings above 35,000 LKR (β = 1.02; S.E = 0.246; p = 0.000). Surprisingly, living far from the market was associated with increased HDDS (β = 0.026; S.E = 0.008; p = 0.004). Conversely, living further away from the main road reduced the HDDS (β = −0.133; S.E = 0.049; p = 0.007). Growing staples reduced the HDDS (β = −0.395; S.E = 0.174; p = 0.023), whereas growing leafy vegetables increased the diet diversity (β = 0.394; S.E = 0.154; p = 0.010). Selling homegrown products also increased HDDS (β = 0.276; S.E = 0.136; p = 0.043). However, other covariates such as the education level of the female adult, household food security status, home garden yield (kg), and livestock richness, which showed significant correlation in the bivariate analysis did not significant in the multiple regression analysis. Although all households in this district engage in some form of home gardening, 79.6% of households did not have adequate dietary diversity. There is a need to understand how home gardens can better contribute to dietary diversity.