Mature oilfield development requires the joint optimization of new well geometry with existing legacy well networks subject to operational constraints. However, existing manual workflows are time-consuming, subjective, and difficult to reproduce. This research presents an automated, rule-aware optimization framework that integrates domain expertise with computational geometry. The framework employs three solvers—differential evolution (DE), particle swarm optimization (PSO), and sequential least squares programming (SLSQP)—to generate reproducible and auditable well-pattern designs. Key innovations include: (i) a grouped geometric transformation scheme that enumerates baseline five-spot layouts with minimal redundancy; (ii) K-dimensional tree (KD-tree) spatial indexing for rapid and scalable matching of thousands of legacy wells to candidate vertices; and (iii) the explicit formalization of expert rules into a constrained multi-objective function with exact penalty methods. A standardized evaluation protocol is used to compare DE, PSO, and SLSQP across different spacing regimes, measuring solution quality, computational efficiency, and robustness over multiple random seeds. Results show that when applied to a large onshore oilfield, the framework reduced baseline plan generation from more than two months to less than 1 h, with each optimization scenario converging within minutes. DE exhibits superior robustness and pattern uniformity. SLSQP achieves the fastest convergence in smooth, lightly constrained landscapes but frequently stalls in local minima under complex regimes, leaving residual violations. PSO offers intermediate performance at a higher and more variable computational cost. The proposed framework serves as a field-deployable tool that substantially reduces the engineering design cycle, improves pattern quality, and enhances the reuse of existing infrastructure.