Climate change affects the established crop production practices by altering rainfall, temperature, and solar radiation—leading to reduced resource use and economic efficiencies. Pertaining to this, we analysed the optimum fertiliser-Nitrogen (N) recommendation of rice and wheat crops for Kharagpur station, West Bengal, India under projected climate change conditions by using six mathematical models, namely, quadratic, square root, linear plateau, quadratic plateau, square root plateau, and the Mitscherlich model. Models were parameterised using field experiment observations at Kharagpur and tested for historical crop yield. The historical (1976–2005) and future (2006–2100) yield responses of rice and wheat to the different N doses are simulated by using a calibrated and validated process-based crop simulation model—Crop Environment and Resource Synthesis (CERES). Eight bias-corrected Global Climate Model (GCM) outputs for the four Representative Concentration Pathways (RCP2.6, RCP4.5, RCP6.0, and RCP8.5) were used to simulate the yield of both crops for the period of 1976 to 2100. Results revealed that the quadratic plateau model represents crop yield response to fertiliser-N application better than other mathematical models. The estimated optimum N dose (OND) increased from an average historical 121 kg N/ha to 138, 143, and 146 kg N/ha for rice, and from an average historical 98 kg N/ha to 119, 124, and 127 kg N/ha for wheat in three future times, respectively, 2006–2035, 2036–2065, and 2066–2095 under all RCPs scenarios. On the other hand, the yield of both crops is expected to decrease in the future due to an increase in N losses, leading to reduced N use efficiency.