Long-term supply of synthetic nitrogen (N) has the potential to affect the soil N processes. This study aimed to (i) establish N response curves to find the best balance between inputs and outputs of N over four ratoons; (ii) use 15N-labeled fertilizer to estimate the N recovery efficiency of fertilizer applied in the current season as affected by the N management in the previous three years. Nitrogen rates (control, 60, 120, and 180 kg ha−1 N) were applied annually in the same plots after the 1st, 2nd, 3rd, and 4th sugarcane cycles. Sugarcane yield, N uptake, and N balance were evaluated. In the final season, 100 kg ha−1 of 15N was also applied in the microplots to evaluate the effect of previous N fertilization on N derived from fertilizer (NDF) and N derived from soil (NDS). Sugarcane yields increased linearly with the N rates over the four sugarcane-cycles. The best balance between the input of N through fertilizer and N removal by stalks was 90 kg ha−1 N in both the 1st and 2nd ratoons, and 71 kg ha−1 N in both the 3rd and 4th ratoons. Long-term application of N reduced NDF from 41 to 30 kg ha−1 and increased NDS from 160 to 180 kg ha−1 N. A key finding is that long-term N fertilization has the potential to affect soil N processes by increasing the contribution of soil N and reducing the contribution of N from fertilizer.