Cluster farming has emerged as a strategic intervention in Ethiopia aimed at fostering sustainable agricultural intensification, enhancing smallholder commercialization, and improving rural livelihoods. Despite strong policy emphasis, its adoption remains low, with limited empirical evidence regarding its broader impacts and the constraints farmers face. This study provides a comprehensive analysis of the determinants and multifaceted outcomes of cluster farming by synthesizing findings from a survey of 421 randomly drawn households (199 adopters and 222 non-adopters), focus group discussions, and key informant interviews. By employing various econometric models including the Heckman selection model, Beta regression, Endogenous Switching Regression (ESR), Entropy Balancing Estimation (EBE), and Augmented Inverse Probability Weighting (AIPW) estimator, the study examines the determinants of farmers’ cluster farming adoption decision, the intensity of their adoption, and the impact of cluster farming on farmers’ commercialization, multidimensional poverty reduction, and food security improvement. The results highlight that cluster farming adoption is significantly influenced by household characteristics such as gender, education, landholding size and fragmentation, access to training and extension services, proximity to main road and market center, and peer influence, while the intensity of adoption is affected by gender, farm size, household size, extension service and off-farm and/or non-farm participation. The findings reveal that the adoption of cluster farming increases farmers’ commercialization, reduces multidimensional poverty, and enhances food security. Specifically, cluster farming adopters earned higher incomes and sold a larger share of their crop production, averaging about $1,106 in crop sales compared to $731 for non-adopters, a difference statistically significant at the 1% level. Moreover, adopter households sold, on average, 50% of their total crop production, compared to 36% for non-adopters, indicating that adopters are more commercial oriented than non-adopters. Crop-specific analysis further supports this trend. Wheat cluster adopters earned $908 and sold 64% of their yield, while non-adopters earned $496 and sold 44%. Similar patterns are observed for maize and teff: cluster farming adopters earned $706 and sold 63% of their maize, compared to $392 and 44% for non-adopters. For teff, adopters earned $942 and sold 68%, while non-adopters earned $386 and sold 45. In terms of poverty, while 77.9% of non-adopters were identified as multidimensionally poor, this figure dropped to 30% among adopters. Likewise, cluster farming demonstrates a substantial positive impact on multidimensional food security, particularly in the dimensions of food quantity, acceptability, and stability. Although the quality dimension showed no statistically significant improvement, the overall food security status of cluster farmers was notably higher than that of non-adopters. These findings underscore the transformative potential of cluster farming in improving rural livelihoods. Therefore, scaling up this farming approach is recommended. To enhance the adoption of cluster farming and its intensity of adoption, the study suggests the need to address gender disparities, strengthen training and extension services, and expand opportunities for off/non-farm activities. Additionally, improving access to farmland, perhaps, through better land lease mechanisms and ownership rights, could encourage greater adoption. Facilitating voluntary land consolidation can also reduce fragmentation and improve access to farmland for cluster farming adoption.