Continuous cultivation without the addition of adequate external nutrient inputs is common in many small-holder farms of Trans Nzoia District. An on-farm trial with a major goal of introducing legume intercrops and green manures in a maize based cropping system to increase crop yields, maintain and sustain soil fertility was started in 1997. A randornised complete block design was used and each farmer served as a replicate. Maize hybrid H614 was intercropped with either common bean (Phaseolus spp.), soybean (Glycine max), groundnuts (Arachis hypogea), and cowpea (Vigna unguiculata). After harvesting the food beans, the four plots were relayed with green manure legumes, sunhemp (Crololaria brevidens), velvet bean (Mucuna spp.) or dolichos (Lablab purpureus). All the three green manure legumes produced low and varying amounts of dry matter (DM) in both seasons. There were significant differences in maize yields following tlie various green manures after one year of incorporation. Maize after Mucuna gave the highest yield of 9.3 tons ha.l followed by maize after Dolichos (7.3 tons ha.l) and lastly that following Crotalaria at 6.7 tons ha.l. These yields were not significantly different from maize fertilized with recommended fertilizer (60 PzOs + 60 N kg ha'l) which yielded 8.5 tons ha,l. Maize yields following green manures and the recommended fertilizer rates gave yields that were significantly (P0.05) higher than the control (5.4 tons ha,l). The yields of maize after two years of green manure incorporation followed a similar trend one year. However maize after Dolichos gave the highest yield of 8.5 tons ha'i. This was followed by fertilized maize (7.4 tons ha.l) and maize following Crotalaria and Mucuna which gave 6.3 tons ha'l each. These yields were also significantly different from that of control plots which gave 5.4 tons ha.l. The utilization workshops for soybean and groundnuts were well attended and even though some farmers insisted on growing them as intercrops due to land pressure, others requested for seed to plant pure stands. The results are promising and may in future add onto the basket of maize production options for small holder maize growers who normally have insufficient funds to purchase recommended inputs for optimum yields.