Background. Undernutrition in early childhood has irreversible and long-lasting implications. Hence, this study was aimed at assessing risk factors of child undernutrition. Methods. A community-based cross-sectional study was conducted on 642 households with mothers to children pairs aged 6–59 months selected by a multistage systematic random sampling method. Child anthropometric measurements on weight were recorded using standardized and calibrated weighing scales. Weight-for-age was compared to the 2007 WHO growth reference by WHO Anthro software. Data were entered using Epi-Info and analyzed using SPSS. Bivariate and multivariate logistic regression analyses were used to evaluate the association between underweight children and their predictors; both crude and adjusted odds ratios with 95% confidence interval were reported. Results. One-fourth (25%) of the children were underweight. Child age (AOR: 2.36), gender (AOR: 1.82), illness (AOR: 0.09), maternal decision making power (AOR: 0.07), maternal education (AOR: 0.19), employment/occupation (AOR: 5.29), and household income (AOR: 4.16) were found to be independent and significant predictors of underweight children. Conclusion. Significant proportion of the children were underweight. Maternal decision-making power persists as a strong predictor of children’s weight. Therefore, intervention programs focusing on improving mothers’ decision-making power on child nutrition would contribute to the efforts towards alleviating the problem.