Alley cropping is a prototype agro-forestry system practiced in many parts of the humid tropics. It is one of the established, promising and sustainable low-input soil management ventures. An improved understanding of the eco-physiological relationships between the arable and perennials in this system could contribute to solving issues of sustainability for nutrient use and system productivity. Field experiments were conducted in Alabata and Ajibode, located in the derived savannah zone of Nigeria, to evaluate the comparative effects of an alley cropping system, arbuscular mycorrhizal fungi inoculation and cassava intercropping on maize yield. These on-farms trials (research managed and farmer managed) were conducted in the 1993/94 and 1995/96 growing seasons, in Rhodic Kandiustalf soil type with low nutrients. Each trial was a split-plot factorial arranged in randomized complete blocks with three replications. Biomass productions of maize and cassava tuber yield were determined. The results obtained at Alabata indicated that the maize plants were heavily infected with AM fungi whether inoculated or not. The hedgerow trees did not have a significant effect on maize yield. The increase in yield brought about by the alley cropping system was generally less than 15%, except in one farm where increases of 34·6 and 46·5% were recorded in AM-inoculated and non-inoculated plots respectively. The introduced AM fungi did not significantly enhance maize yield at Alabata; the percentage increase was not more than 24%. This trend was also observed at Ajibode. Cassava tuber yield was found to be suppressed by the alley cropping system irrespective of the cultivar. The decrease in yield ranged from 2·6 to 150·8%. However, inoculation of the cassava with exotic AM fungi brought about an enhancement of its tuber production. The increase ranged from 20·7 to 189·8% depending on the treatment combination. It is therefore necessary to encourage adoption of the improved alley cropping system integrated with mycorrhizal technology.