With the goal of minimizing the maximum makespan of Flexible Job-shop Scheduling Problem (FJSP), a hybrid growth optimizer considering multiple critical paths was proposed based on the Growth Optimizer (GO). First, a two-stage coding mechanism based on process sequence and machine allocation was proposed, enabling the growth optimizer to be applied to the solution of FJSP. Second, a heuristic method was applied to improve the quality of the initial population. An improved reflection strategy and a multi-critical path optimization strategy were designed for local search to accelerate algorithm convergence and enhance the algorithm’s ability to escape from local optimal solutions. Furthermore, a tabu search algorithm is introduced in the later stages of algorithm convergence to further enhance the algorithm’s ability to escape local optima. Finally, a large number of simulation experiments were conducted using 10 benchmark examples and a processing example. Experimental results show that the HGO algorithm improves the average optimization efficiency by 4.53% compared to the suboptimal algorithm in the Brandimarte benchmark, proving the effectiveness and superiority of the HGO algorithm in solving FJSP.