A long-term (1982 to 2000) field experiment was conducted at Zhangye, Gansu, China, on a sandy clay loam (Typic Anthrosol) under wheat (Triticum aestivum L.)–wheat–corn (Zea mays L.) rotation to determine the effects of N, P, and K chemical fertilizers and farmyard manure (M) on grain and straw yield, harvest index (HI), protein concentration, and N uptake in grain and straw and accumulation of nitrate N (NO3–N) in the soil profile (0–180 cm). The eight treatments from various combinations of fertilizers and M were check, N, NP, NPK, M, MN, MNP, and MNPK. Mean grain yield decreased in the order of MNPK ≥ MNP > NPK > MN > NP > M > N > check (i.e., 8.01, 8.00, 7.51, 7.28, 7.00, 5.50, 4.89, and 3.43 Mg ha−1, respectively). Yield response to applied N and P increased with time since yields in the check plots declined with time. Potassium fertilizer application provided no, slight, and dramatic increase in grain yield during the initial 6 yr, next 5 yr, and last 8 yr, respectively. Response of straw yield to fertilizers and M was similar to the grain yield. Mean HI increased with fertilizers in no-M treatments for both crops. Crude protein concentration and N uptake in grain and straw increased markedly with fertilizers, and M increased it further. Fertilizers (N, NP, and NPK) led to NO3–N accumulation in most subsoil layers. Combined applications of fertilizers and M reduced soil NO3–N accumulation in soil compared with fertilizers alone. In conclusion, the findings suggest that it is important to use balanced application of chemical fertilizers and M at proper rates in order to protect soil and underground water from potential NO3–N pollution while also sustaining high crop production.