When dealing with the state estimation of an unknown but bounded system, existing particle filter methods generally have problems of high particle demand and particle degeneracy, which affects the accuracy of state estimation. This paper designs an adaptive firefly based-interval particle filter. Firstly, the weight of each interval is calculated through its width and estimation error. Then, an adaptive firefly optimization strategy is introduced in the resampling step. By calculating the optimized adaptive coefficient, the moving direction and step size of each particle interval are determined to improve the posterior particle interval distribution. In addition, the state estimation interval is further divided, and the upper and lower bounds of the obtained state estimation are iteratively contracted to obtain smaller state estimation interval boundaries and more accurate state estimation results. The proposed algorithm can make intervals with higher weight coefficients wrap the true state more effectively, thereby reducing the particle demand. Moreover, the designed adaptive resampling strategy can significantly reduce the degree of particle degradation. Finally, through numerical simulation and experiments on the Buck-Boost model, it is verified that the proposed algorithm can wrap the upper and lower bounds of the state estimation more tightly and has a lower root mean square error, indicating that the designed filter improves the accuracy of state estimation and provides a tighter state wrapping.