Coastal wetland vegetation is at risk of severe degradation at multiple scales due to multiple stresses. Although numerous studies have emphasized the importance of scale dependence, few studies have quantitatively explored how temporal-spatial scales affect vegetation coverage (VC) in coastal wetlands. To identify the effect path of scale on the driving mechanisms of VC, we constructed a structural equation model (SEM) to quantify the temporal-spatial scale dependence of the driving mechanisms of VC in coastal wetlands in China at three spatial scales (national, regional and local) mainly over a 19-year period (2000–2018, including the initial, middle and later stages of vegetation succession). The results showed that at all spatial scales, natural factors were the strongest drivers of VC at the later stage of succession (14–39 a). With reduced spatial scales, soil and topography (only middle stage, 9–11 a) played more important roles at the middle and later stages, whereas human factors had the opposite effect. The common medial factors of the driving mechanisms underlying VC were topography and soil in all scale models. The across-scale model explained 21.8% and 40.4% of the elevation and pH, respectively. The temporal-spatial scale dependences of topography and soil were stronger than those of other variables. Our findings highlight that temporal-spatial scales mainly controlled the driving mechanisms underlying VC via topography and soil in coastal wetlands. This study is the first to quantitatively reveal the effect path of scale on the driving mechanisms underlying VC in coastal wetlands.