Measurements of vegetation are most commonly used in evaluations of restoration projects, with less frequent analysis of soils, fauna, and hydrologic characteristics. Although particular characteristics of projects, such as vegetative cover and production, can resemble those in similar naturally occurring wetlands, overall functional equivalency has not been demonstrated. However, ongoing research is providing information on what can and cannot be accomplished, valuable insights on how to correct mistakes, and new approaches to defining success. The challenge is how to recognize and deal with the uncertainty, given that projects are ecologically young and that our knowledge of the process of restoration is evolving. One way to deal with the uncertainty is to use scientific principles of hypothesis testing and model building in an adaptive management framework. In this way, options can be systematically evaluated and needs for corrective actions identified when a project is not progressing toward goals. By taking such an approach we can improve our ability to reliably restore wetlands while contributing to our understanding of the basic structure and function of ecosystems.