This paper proposes ProgressEditor that performs high-quality diffusion distillation-guided 3D scene editing in a novel progressive manner with a simple yet effective framework. Inspired by the crucial observation that the multi-view inconsistency issue in the scene editing task is rooted in the diffusion model's large feasible output space (FOS), we introduce a novel framework to control the size of FOS and reduce inconsistency by decomposing the full editing tasks into several subtasks, and progressively perform each of them on the scene. Within this framework, we design a difficulty-aware subtask decomposition scheduler and an adaptive 3D Gaussian splatting (3DGS) training strategy, to perform each decomposed subtasks in high efficiency. Extensive evaluations show that our ProgressEditor produces state-of-the-art scene editing results in various scenes and challenging editing tasks, through a simple framework without any expensive or sophisticated add-ons like distillation losses, components, or training procedures. Notably, ProgressEditor also provides a new way to preview, control, and select the aggressivity of editing operation during the editing process.