Recent empirical studies on farm mechanization highlight its significant role in transforming the livelihoods of rural smallholder farmers. Mechanization enhances productivity and efficiency, reduces post-harvest losses, enables deeper ploughing, and allows for more precise input application. It also reduces labor intensity, making it possible for households to reallocate labor toward education and non-farm activities, thereby improving overall household welfare. Nevertheless, mechanization may negatively affect farm diversification and nutritional security, issues often underrepresented in policy discourse. In response to these complexities, this study was initiated to investigate the determinants and impacts of mechanization on production efficiency, multidimensional rural poverty, and food and nutrition security in Ethiopia’s Arsi Zone. To achieve this, data were gathered from 413 randomly selected households between June and August 2023. Primary data were collected through household surveys, focus group discussions, and key informant interviews, while secondary data were sourced from relevant institutions. The analysis employed a range of quantitative models, including multivariate probit, Tobit, ESR, CESR, ESRpo, IV fractional probit, PSM, and MESR, to ensure robust and comprehensive insights. Descriptive analysis revealed that 59% of households use combine harvesters, 51% use tractors, and 40% adopt both, with labor shortages and productivity gains cited as primary drivers. According to the multivariate probit and Tobit models, education, access to credit, and cooperative membership significantly increased both adoption and intensity of mechanization, whereas land fragmentation and crop diversification negatively influenced uptake. The MESR model demonstrated that mechanization, whether adopted individually or jointly, substantially improved wheat technical efficiency, boosted productivity, and reduced poverty. Moreover, ESR results indicated a 70% increase in the likelihood of higher income for adopters, while CESR and ESRpo models revealed a 43% increase in food security likelihood and a 23.86% reduction in HFIAS scores, thereby affirming mechanization’s role in supporting Sustainable Development Goals such as “No Poverty” and “Zero Hunger.” However, results from the IV fractional probit and PSM models showed that mechanization significantly reduced agricultural diversification among mechanized farmers, as indicated by PSM result declines in the Agricultural Diversification Index (−0.0993), Crop Diversification Index (−0.119), and Livestock Diversification Index (−0.0789), which has important implications for dietary diversity. Based on Multitreatment Endogenous Switching Regression, diversification in the absence of mechanization improved dietary diversity (HDDS) by 0.761 units (t = 4.15) and food consumption scores (FCS) by 5.60 units (t = 12.60), while diversification in mechanized households increased HDDS by 0.65 units and FCS by 1.41 units, both significant and positive. However, mechanization improved food access for non-diversified households (FCS + 1.15, t = 2.224) but reduced dietary diversity (HDDS − 0.609, t = −5.985) and food consumption scores (FCS − 1.24, t = −4.16) among diversified households, indicating a potential trade-off between mechanization and nutrition. Based on these findings, the study concludes that mechanization adoption is a multifaceted process shaped by household assets and technology attributes. While mechanization improved rural households’ welfare, its interaction with diversification and nutrition warrants careful policy consideration. Mechanization policies should be accompanied by complementary measures that expand credit access, strengthen cooperatives, and improve extension services, while addressing potential adverse effects on diversification and nutrition.