Buildings account for a substantial share of global energy consumption and greenhouse gas emissions, creating an urgent need for design strategies that simultaneously address operational performance, occupant comfort, and life-cycle environmental impacts. While simulation-based optimization has become increasingly common in building performance research, relatively few studies evaluate energy use, thermal comfort, and embodied carbon within a unified tri-objective framework. This study presents a simulation-based tri-objective Pareto optimization of residential buildings in Riyadh, Saudi Arabia, and Dubai, United Arab Emirates, using DesignBuilder, EnergyPlus, and the Non-Dominated Sorting Genetic Algorithm II (NSGA-II). A standardized four-story residential apartment prototype comprising 16 thermal zones and 2239.82 m² of conditioned floor area was developed and simulated under identical geometric, operational, and HVAC assumptions. Window-to-wall ratio, glazing type, external shading depth, and cooling setpoint temperature were optimized to minimize annual site energy consumption, ASHRAE 55 thermal discomfort hours, and embodied carbon emissions. Baseline simulations revealed substantially higher operational demand in Dubai, with annual energy consumption reaching 272,077 kWh compared with 196,478 kWh in Riyadh, while discomfort hours increased from 2,530 h/year to 3,262 h/year. Optimization reduced annual energy demand by 72.9% in Riyadh and 74.5% in Dubai, while thermal discomfort was reduced to 776 h/year in the best-performing comfort solution. Pareto-optimal solutions consistently favored low window-to-wall ratios (10–16%), high-performance glazing, and external overhangs between 1.5 and 2.0 m. The findings demonstrate the effectiveness of tri-objective optimization for balancing operational efficiency, occupant comfort, and embodied carbon while providing climate-responsive façade design guidance for residential buildings in hot-arid Gulf environments.