This study contributes to understanding the relationship between climatic variables and groundnut production in different farming systems in Uganda. Alternative production function models are estimated using pooled cross-sectional time series data at the district level. The models incorporate land area, indicators for farming systems, technological change, and either rainfall or the El Niño–Southern Oscillation (ENSO) effect as variables to account for climatic conditions. The data set includes 333 observations corresponding to 37 districts for 9 consecutive years, from 1992 to 2000. Analyses were performed using a Translog functional form and GARCH estimators. The results suggest that the partial elasticity of production for land is positive, high and significant, which is consistent with a priori expectations. Farming systems are also found to have a significant impact on output variability. Climatic conditions, measured by rainfall, have a non-significant effect; but, when the ENSO phenomenon is used instead a significant negative effect is detected particularly for the warm phase. An important and alarming finding is a marked negative rate of technological change revealing productivity losses over the time period studied.