The pressing need to increase food production in the tropics to feed the burgeoning population of the Third World requires that crop yields per ha must be increased without prejudicing the resource base for future generations. Biological nitrogen fixation (BNF), especially that associated with legumes, has great potential to contribute to productive and sustainable agricultural systems for the tropics, but more research is required to investigate how biologically fixed N, and the increased BNF contributions resulting from research innovations, can be incorporated into viable agricultural systems to increase crop or pasture yields and to substitute N fertilizer inputs. A majority of the soils of the humid and semi-humid tropics have mineral fractions composed of 1: 1 lattice clays or sesquioxides of relatively low capacity to retain nutrients (CEC) and water (WHC). It is the soil organic matter (SOM) which has high CEC (after liming) and WHC, and soils under undisturbed climax vegetation are usually high in organic matter which is responsible for their fertility. The key to the long term fertility of such soils is to maintain their soil organic matter by the preservation of crop residues and the selection of suitable crop rotations or fallows. In this review we examine several types of agricultural systems utilized in the tropics ranging from pastures, ley cropping, zero-til1 rotations as wel1 as green manuring and discuss the management options that can be adopted to preserve their agricultural productivity through the strategic use of legumes in these systems, and their effects on pasture and crop yields. The introduction of forage legumes into tropical pastures can increase and sustain their productivity, with only modest inputs of lime and P and K fertilizers. Similarly, crop and pasture rotations (ley cropping) maintain SOM and soil fertility and crop yields can benefit greatly from the introduction of pasture legumes into the ley. Continuous cropping under minimum or zero tillage can maintain soil cover, and stimulate the retention of SOM, such that nutrient losses are often minimal, and legume derived N can be efficiently transferred to subsequent crops. The options for the resource-poor small-holder to efficiently utilize biologically fixed N as a N supply for cereal grains are more limited and need more attention from researchers as wel1 as less neglect from government organizations. The addition of lime and P fertilizer in modest quantities in many under-developed regions could make large contributions to increased crop yields. If such modest fertilizer inputs were to be combined with suitable crop rotations including green manure or grain legume crops, larger increases in crops yields could be achieved on a sustainable basis, but in many regions agricultural extension services are non-existent and poor farmers have little access to even these basic chemical inputs.