This study was conducted to assess the soil and air temperature levels of dominant tree species in agroforestry systems that influence Coffea arabica productivity. A purposive sampling technique was employed to select villages and farms from different agroecological zones. The data collection methods included questionnaires and focus groups with key informants. Soil and air temperatures were measured using temperature sensors. General linear models in SAS were employed to analyze the temperature data and Statistical Package for Social Science for socio economic data. Dominant tree species included Grevillea robusta, Albizia schimperiana and Rauvolfia caffra. There was a significant difference in soil and air temperature regulation among tree species in the midland (p < 0.05), with mean temperature differences of 0.5–1.6 and 0.2–0.4 °C for soil and air temperature, respectively. G. robusta significantly regulates soil and air temperature in both highland and midland zones (p < 0.05) compared to the other studied tree species that had mean temperature differences of 0.2–1.6 and 0.3–0.4 °C for soil and air temperature, respectively. Since moderate temperature favours Coffee productivity, G. robusta is recommended in both the midland and highlands; however, it is imperative to investigate how the soil impacted by G. robusta affects coffee productivity.