Worldwide wetland loss over the past 50 years has made wetland conservation a public policy priority, leading to an increase in wetland restoration programs. However, predicting long‐term restoration outcomes remains difficult. The monitoring of these programs rarely exceeds 5–10 years, forcing wetland managers to rely on short‐term success criteria that may be criticized by the scientific community. Our objective was to assess the significance of four short‐term success criteria ( Carex ssp. shoot density, Salix ssp. survival, invasive species cover, and hydrologic dissimilarity to reference sites) used in a restoration program of 12 wetlands monitored for 5 years post‐restoration in predicting restoration outcomes 15 years post‐restoration. We defined the success of restoration efforts after 15 years using a cluster analysis‐based approach, and the clusters were described using principal coordinate analysis and Tukey's post hoc honest significant difference test. Finally, we assessed the pertinence of each short‐term success criteria in predicting long‐term restoration outcomes using Pearson correlation tests and spatial regressive models. Our results demonstrate that stress‐based short‐term success criteria can be reliable predictors of longer‐term success for communities with shallow water tables, whereas target‐species‐based short‐term success criteria are not. Hydrologic dissimilarity to the reference site was appropriate for willow‐sedge community outcome predictions, while invasive species cover was best for sedge community outcome predictions. For communities in drier habitats, such as the willow‐herb community, none of the tested short‐term success criteria were significant predictors of long‐term restoration outcomes, and further research is required to identify suitable short‐term success criteria.