The central mid-hills region of Nepal is characterized by rainfed production systems where maize (Zeamays) is the principal crop during the monsoon (kharif) season. In general, maize yields in the hillsare judged to be intractably low and few efforts have systematically assessed either the water-limitedproductivity potential or identified sensible entry points toward sustainable intensification that can beselectively matched to the needs and constraints faced by different types of farmers. This study combinesfield surveys, on-farm field experiments, and simulation modeling (DSSAT) to explore opportunitiesfor closing maize yield gaps in these ecologies. Among surveyed households, the mean grain yield ofmaize under current farmer practice was 2.0 t ha−1whereas good agronomic practices increased maizeyields up to 6.5 t ha−1(i.e. exploitable yield gap of 4.5 t ha−1). Recognizing that farmers rarely adoptfull technology packages, the value of single agronomic interventions was also explored with all othermanagement factors maintained as per prevailing farmer practice. Averaged across sites and two seasons,non-limiting fertilizer rates (175:60:60 kg NPK ha−1) increased grain yields by 1.8 t ha−1followed by theuse of hybrids (1.4 t ha−1), higher plant population (0.9 t ha−1), and improved weed control (0.9 t ha−1).These results were also reflected in changes to the economic performance of the system with grossmargins increasing from a ‘farmer practice’ base of $202 ha−1(B:C = 1.86) to $339 ha−1(B:C = 2.14) withhybrid adoption, $454 ha−1(B:C = 2.88) with higher plant population, $646 ha−1(B:C = 2.52) with non-limiting fertilizer, and $611 ha−1(B:C = 3.37) with careful weed control. Further gains in profitability wereachieved with layered agronomic interventions, which increased gross margins to $857 and $763 ha−1(B:C = 2.95 in both) in conventional tillage (CT) and minimum (strip tillage; ST) systems, respectively.Divergence between grain yield gains and economic performance criteria highlights the importance ofconsidering both perspectives. For resource-poor households, maintaining optimal plant population canincrease B:C by more than 50% with small investments ($7 ha−1). Simulation results also suggest thatadditional potential productivity advantages are achievable with timely planting (e.g. 19% higher meanyield for short duration hybrid with less inter-annual variability). Cultivation of longer duration hybridscan also increase yield potential over their shorter duration alternatives, although trade-offs with yieldpotential of the second crop in the rotation must be considered. These results highlight several pathwaystoward intensification in the hills of Nepal that have varying investment requirements, inferred risks,and implications for the different dimensions of sustainability and food security.