The study investigated the nutritional status of under-five children of farm households. The study utilized primary data from 352 farm households with 140 under-five children. Household crop commercialization index (CCI) was used to estimate cassava farm household crop sale ratio and categorize the households into four commercialization levels while WHO Anthro software was employed to analyze under-five children anthropometric indices such as weight-for-age z-score (WAZ), height-for-age z-score (HAZ) and weight-for-height z-score (WHZ). Logit regression model (LRM) was used to examine the drivers of under-five children’s nutritional status of farm households. The study found that 42.9%, 7.9% and 3.6% of the children are stunted, underweight and wasted respectively. The highest stunting level was recorded in zero level households (CCI 1). Although, some higher CCI households (medium-high and very-high level) recorded increased percent of stunted children. This revealed that being a member of low or high-level commercialization households may not guarantee better nutritional status of young children of farm households. The results of LRM indicated that the predictors of children nutritional status were child’s age, farm size, access to electricity, healthcare and commercialization variables. Moreover, weak positive and negative relationships exist between CCI and children’s nutrition outcomes as measured by the z-scores. The study recommended maternal nutrition-sensitive education intervention that can improve nutrition knowledge of mothers and provision of infrastructure that enhance increased farm production and promote healthy living among farm households.