Pakistan faces significant future challenges to feed a growing population, with 47 % of the populace currently categorised as food insecure and water becoming scarcer. The current rice-wheat (RW) cropping system is critical for food security, but exhibits low productivity due to sub-optimal varieties, timings, and nitrogen (N) stress. We investigated farmer management strategies to increase the productivity of the RW system while quantifying trade-offs with environmental sustainability (particularly nitrate leaching). Cropping systems simulation with a validated model (APSIM) enabled analysis of a broad range of current management practice modifications (crop sowing dates, varieties, fertiliser rates, and deficit irrigation strategies) in a large Monte-Carlo Matrix (13,000+ elements). We evaluated annual crop production, gross margins (GMs), irrigated water productivity (WPi), nitrate leaching, ET and the crop transpiration proportion of total ET, for all combinations on an annual basis over a 35-year simulation period (1980−2014). We found that WPi in both the highest (Gujranwala) and lowest yielding (Narowal) districts of Punjab can be maximised by using a shorter-season rice cultivar sown slightly later than current farmer practice and irrigated under alternate wetting-and-drying (AWD – with 2–4 days gap), with increased N fertiliser. The currently-used wheat variety was found to be satisfactory and capable of capitalising on the earlier sowing possible through timelier rice harvest, yet with a more generous wheat irrigation strategy (4−5 irrigations rather than 3−4) and N fertiliser regime. The currently-used long-season rice variety maximises GM at Gujranwala, largely due to premium grain prices for that variety, however in this water-limited cropping environment maximising WPi will maximise regional production. If no N-leaching limits apply, system WPi at Gujranwala can be increased by a factor of 2.67x over current farmer levels, via the indicated management changes, with associated N fertiliser increases of 1.8–2.2 times at Gujranwala. If annual N leaching limits are imposed (50, 100 or 150 kg N ha−1), those achievable WPi gains change to 2.0x, 2.6x, and 2.67x respectively at Gujranwala, with limits on applied N fertiliser being the only changes in optimised agronomic management. A key finding was that even under the strictest N-leaching limit (50 kg N ha−1 yr−1, which is similar to what farmers currently lose), it is still possible to significantly increase system water productivity (by 2.0x) through the optimised agronomic changes, while using less fertiliser N (0.8x). Our research indicates considerable gains in rice and wheat food production are possible for the Pakistani Punjab through farmer management changes, however risks associated with uncertain water supplies need to be assessed.