Farm efficiency analysis provides significant insights into farms’ potential to enhance agricultural productivity. This article reports on an investigation of technology adoption and technical efficiency (TE) in the Ethiopian maize sector. We estimated TE while accounting for the potential technological difference between improved and local maize varieties and addressing self-selection bias resulting from farmers’ decisions to adopt new crop varieties. Using comprehensive household-level data collected in 2011 from five major maize-producing regions in Ethiopia, we specified a stochastic frontier model to estimate TE and employ propensity score-matching technique to address self-selection bias. The result confirm that imposing a homogenous technology assumption for improved and local maize varieties biases efficiency estimates and the ranking of farmers based on their efficiency scores. The mean TE of 66.18%, estimated after correcting for technology difference and self-selection bias, indicated that an increase of around 33.82% in maize productivity could be achievable with the current input levels and technology.