Water scarcity in semi-arid regions threatens agricultural productivity, groundwater reserves, and rural livelihoods. Conventional irrigation planning often relies on single-objective optimization, which fails to capture the multi-dimensional trade-offs among water conservation, economic returns, crop diversity, and aquifer health. This study develops a 4D multi-objective optimization (MOO) framework for irrigation water resource planning and applies it to wine grape production at the Chinangali Irrigation Scheme in Dodoma, Tanzania. The model simultaneously minimizes seasonal irrigation water demand while maximizing net profit, crop diversification, and groundwater sustainability. Crop water requirements were determined using the FAO Penman–Monteith method across four grapevine varieties: Chardonnay, Cabernet Sauvignon, Chenin Blanc, and Riesling. A hybrid decision-support approach integrating Weighted-Sum linear programming with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) generated 47 non-dominated Pareto-optimal solutions. The results demonstrate the structural limitations of single-objective strategies: profit maximization yields maximum net returns (TZS 2,098 million) but drives groundwater extraction to an unsustainable level (Groundwater Sustainability Index, GSI = 4.81%), whereas water-only minimization severely depresses farm revenue (TZS 1,087 million). A balanced, equal-weight compromise solution prescribes an optimal allocation of 18.5 ha Chardonnay, 31.2 ha Cabernet Sauvignon, 27.5 ha Chenin Blanc, and 42.8 ha Riesling. This strategy achieves TZS 1,876 million in net profit (89.3% of peak profit), saves 31,490 m3 of water annually relative to current practices, and raises the GSI to 12.43%. Post-optimal sensitivity analysis confirms the operational stability of this compromise across varying stakeholder priorities. The proposed framework provides a robust tool for negotiating sustainable water management in water-scarce agricultural systems.