There is a strong need to identify agricultural management practices that maintain agronomic productivity while diminishing soil N2O emissions. The yield-scaled N2O emissions (YSNE) indicator can help to evaluate the adequacy of a given agricultural practice under both aspects. Long-term (18-yr) soil water and mineral N dynamics, crop biomass and yields, and 2011-2012 soil N2O emissions and ancillary variables were measured on barley (Hordeum vulgare L.) production in a tillage (conventional tillage, CT; no-tillage, NT) and N rate (0, 60 and 120 kg N ha-1) combination under rainfed Mediterranean conditions (NE Spain). Once evaluated, the STICS soil-crop model was used to simulate the 18-yr soil N2O emissions of each tillage system under increasing N rates (0, 30, 60, 90 and 120 kg N ha-1) in order to identify optimum management to reduce YSNE, being initialized with observed data. Cropping season precipitation was highly variable during the experiment, being a key regulating mechanism for crop yields and simulated soil N2O emissions. Crop yield under NT with N outperformed CT in 11 years. STICS performed reasonably well when simulating cumulative N2O emissions and ancillary variables with model efficiencies greater than 0.5. The 18-yr average simulated cumulative N2O emissions were 0.50, 0.82 and 1.09 kg N2O-N ha-1 yr-1 for CT-0, CT-60 and CT-120, respectively, and they were 0.53, 0.92 and 1.19 kg N2O-N ha-1 yr-1 for their counterparts under NT. These averages mask a large variability between years, according to precipitation. The 18-yr mean yield-scaled N2O emissions were 2.8 to 3.3 times lower under NT, compared to the corresponding CT treatments. Under CT, N application would increase YSNE in most years while YSNE would be more resilient to the application of increasing N rates under NT. Our work demonstrates that in rainfed Mediterranean systems NT is a win-win strategy for the equilibrium between agricultural productivity and low soil N2O emissions.