This study was conducted in Zomba District of Southern Malawi to evaluate the profitability of mixed tree intercropping and relay cropping agroforestry technologies. Thondwe, Malosa and Dzaone Extension Planning Areas of the district agriculture office were selected for the study. A total of 286 farmers was targeted, 74 relay cropping Tephrosia vogelli/candida with maize (RA), 101 practicing mixed tree intercropping of Gliricidia sepium with maize (MA) and 119 non-adopters. The study used primary data which was analyzed using descriptive statistics, Gross Margins (GMs), Benefit-Cost Ratio (BCR), Net Present Value (NPV) and the Expected Variance (E-V) model. Both MA and RA farmers had positive NPV estimated over a period of twenty years at MK 52,418.53 (US$374.42) and MK 10,573.69 (US$75.53) respectively. Non-adopters had negative NPV of MK -7,283.84 (US$52.03). GMs per hectare for MA farmers were the highest. The BCR for the two agroforestry technologies were greater than 1 at 1.6 for MA and 1.12 for RA, implying that it is worth investing in the technologies. The results of the optimization of MA and RA using the E-V programming model showed that if risk is not considered in the optimization of farm resources, farmers can optimize the use of their resources with the production of maize under the two agroforestry technologies only. At the optimal level, the farmers cultivated 1.27 hectares of MA and 1.1 hectares of RA only. When risk was introduced into the optimization, farmers started withdrawing land from maize in agroforestry to maize produced without agroforestry, which reduced their expected income. The results also showed that farmers considered MA to be more risky compared to RA technology.