This paper proposes Forest Growth Optimization (FGO), a novel nature-inspired metaheuristic algorithm for solving continuous global optimization problems. The algorithm is inspired by Forest ecological succession: Pioneer species colonization, canopy closure, climax community emergence, and it incorporates Multi-stage succession (pioneer -> intermediate -> climax), gap dynamics, seed dispersal as its core search mechanism. Unlike existing metaheuristics that rely on a single update rule applied uniformly across the population, FGO introduces a set of named operators that collectively capture the multi-phase dynamics of its biological inspiration. Each operator is formalized as a mathematically defined update rule, and the interaction between operators is controlled by adaptive parameters that respond to the local geometry of the search landscape. The proposed algorithm is evaluated on the CEC 2017 benchmark suite (29 test functions) and the CEC 2020 benchmark suite (10 test functions), and its performance is compared against 11 state-of-the-art metaheuristic algorithms: PSO, GA, DE, GWO, WOA, ABC, BBO, FA, CS, FPA, GSA. The experimental results demonstrate that FGO achieves the best mean fitness on 24 out of 29 CEC 2017 functions and on 8 out of 10 CEC 2020 functions. The Friedman test ranks FGO first with an average rank of 1.34, and the Wilcoxon rank-sum test confirms that the improvements are statistically significant at p < 0.05 on at least 22 of the 29 CEC 2017 functions. An ablation study quantifies the contribution of each operator to the algorithm's overall performance, with the most critical operator producing a degradation of up to 27.3% when removed. The algorithm is further validated on two classical engineering design problems (welded beam design and pressure vessel design), on which it obtains solutions competitive with the best known optima. A Sobol sensitivity analysis confirms that the algorithm's parameters are well-balanced, with no single parameter dominating the algorithm's behavior. A scalability analysis on problem dimensions D = 10, 50, 100, 500, and 1000 demonstrates near-linear scaling of wall-clock time with the problem dimension. The results collectively demonstrate that FGO is a competitive metaheuristic for continuous global optimization, with broad applicability to engineering design, Machine Learning (ML) hyperparameter tuning, and scientific computing.