<p>Time, cost, quality, and safety (TCQS) are critical and interdependent objectives in modern construction projects, where increasing complexity necessitates advanced multi-objective optimization techniques. This study proposes an Opposition-Enhanced Arithmetic Optimization Algorithm (OE-AOA) that integrates a simple opposition-based learning (OBL) mechanism to enhance the exploratory capability of the conventional Arithmetic Optimization Algorithm (AOA). By incorporating opposition during both population initialization and iterative updating, the proposed method effectively expands the search space, mitigates premature convergence, and improves solution diversity. A real-world 13-activity building project with multiple execution modes is modeled to capture discrete time–cost trade-offs, quality performance levels, and safety risk ratings, resulting in a complex nonlinear multi-objective optimization problem. The performance of the proposed OE-AOA is evaluated against the classical AOA, Latin Hypercube Sampling (LHS)-based NSGA-III, and the Adaptive Opposition Slime Mold Algorithm (AOSMA). The results demonstrate that OE-AOA generates a denser and more uniformly distributed Pareto front, achieving superior performance in project duration, total cost, overall quality, and cumulative safety risk. Furthermore, the proposed approach exhibits higher convergence accuracy, improved stability, and greater robustness across multiple simulation runs. Overall, the findings confirm that OE-AOA provides a reliable, computationally efficient, and high-performing framework for integrated TCQS optimization, supporting decision-makers in achieving balanced, safe, cost-effective, and high-quality construction project outcomes.</p>