Establishing an accurate model of the combustion system of a Circulating Fluidized Bed (CFB) boiler is particularly important to improve the control performance of the boiler combustion system. Aiming at the problems of too many adjustment parameters and insufficient accuracy in the existing modeling methods, a modeling method based on Lévy Flying Dual Chaos Sparrow Search Algorithm (LF-DCSSA) is proposed, which means the improved tent chaotic mapping and lens imaging inversion are used to initialize the population and enhance the population diversity and quality, meanwhile, introduce Lévy flight strategy, nonlinear adaptive decreasing weights and chaotic perturbation to improve the global search capability, convergence speed and optimization accuracy of the algorithm. The CFB boiler combustion system is modeled by using the LF-DCSSA algorithm and the model validity is verified. The experimental results show that the LF-DCSSA algorithm can establish a more accurate combustion system model, which provides a new way for the rapid identification of the system model.