Soybean yields on smallholder farmers’ fields in Malawi are constrained by poor soil fertility, limited application of external inputs and poor crop husbandry. We tested crop management practices through on-farm experimentation and participatory technology evaluation. Two agronomic soybean trials were established in 2009 and 2010 in three contrasting agro-ecologies resulting in 72 replications per trial. Treatments in the first trial included several combinations of inoculation with Bradyrhizobium japonicum, inorganic fertiliser application and compost manure. In the second trial, farmers tested an improved variety, optimal weeding regime, increased plant population and chemical pest and disease control. A combination of inoculation, inorganic fertiliser (10 N, 8 P, 20 K in kg ha−1), and 6 t ha−1 compost manure increased yields from 0.86 t ha−1 under farmers’ practice to 1.56 t ha−1 and resulted in average profits of US$222 ha−1. Increased plant populations and biocide spraying also resulted in substantial yield increases. Inoculation and increased plant population resulted in an average value to cost ratio (VCR) > 2. Low investment costs make inoculants, compost manure and increased plant populations interesting options, whereas adoption of inorganic fertiliser application in soybean may be limited due to high costs and low VCR. The farmers ranked eight technologies in descending order of preference as: (i) early planting, (ii) plant population, (iii) variety choice, (iv) compost manure, (v) weeding, (vi) inoculant, (vii) fertiliser and (viii) spraying. Our participatory research approach demonstrated that there is a wide range of technologies with different levels of human and financial investment costs that smallholder farmers can adopt to enhance their soybean yields and profits.