Scant rainfall and poor soil fertility are the two major obstructions to crop production on the Loess Plateau. To improve crop productivity and to reduce N fertilizer rates, a 4-yr field experiment was conducted to investigate the effects of leguminous green manure (GM) and N fertilizer on winter wheat (Triticum aestivum L.) growth, yield, and economics on the Loess Plateau. Following a split-plot design, the main treatments included three legume species: Huai bean (Glycine ussuriensis Regel et Maack.), soybean [G. max (L.) Merr.], mung bean (Phaseolus radiatus L.), and summer fallow (as control treatment [CK]); The subtreatments included four N fertilizer rates that were applied to the wheat. Leguminous GM apparently improved wheat growth, productivity, and nutrient uptake compared to bare fallow, especially during a wet year. At least 2 yr and abundant rainfall are required for bettering the GM approaches. Incorporation of GM for 4 yrs could effectively reduce the N fertilizer rate for wheat by 33% (54 kg N ha(-1)), with even more potential during a wet year. High expenditures for field management and variable weather patterns led to few direct economic benefits of GM approaches. Huai bean is a more profitable legume species to be used as GM crops. The cultivation of leguminous GM during summer is a better option than bare fallow for sustaining wheat productivity, and decreasing the required N fertilizer rates not only on the Loess Plateau of China but also in the other similar dryland regions around the world.