Response to fertilizer is a function of several factors including weather, soil condition, previous year yield level and the actual amount of fertilizer applied. The main objectives of this paper were to evaluate if cumulative precipitation (PPT), growing degree days (GDD), soil moisture (SM) and fertilizer can explain variability in long-term grain yield and fertilizer response index (RI); to assess to assess if demand for N fertilizer is greater the ensuing year given low or high yield levels in the previous year; and to assess if rainfall distribution coupled with the previous year grain yield level and RI of winter wheat (Triticum aestivum L.) improves the predictability of ensuing year RI. Data from two long-term experiments located at Stillwater and Lahoma, OK were used in this report. The experimental design was a randomized complete block with four replications. Twelve (Stillwater) and thirteen (Lahoma) treatments that consisted of annual application of N, P and K inorganic fertilizers were included although selected treatments were used to address the objectives. Soil moisture at 75 cm depth was a better predictor of both grain yield (R² =0.35) and RI (R² = 0.66) at Stillwater while SM at the 60 cm depth was significantly correlated to grain yield (R²= 0.39) at Lahoma. For years where the previous year yield was low, the ensuing year had a low RI while in many years where the previous yield was high; the ensuing year had a high RI in both experiments. The practical significance of these findings is that producers could theoretically adjust preplant or in-season topdress N fertilizer rates by considering the previous year yield level and the degree of response to N.