Species richness has been viewed as the primary measure for understanding the biodiversity in a community for a long time. If the research interest focuses on a forest community, it is customary to collect data with a quadrat sampling. However, several well-known estimators cannot provide reliable consequences when the fraction of sampling domain is less than 5%, which is a common situation in practice. To meet the practical requirement, this study presents some promising methods that yield encouraging results in light of the various aspects of statistical criterion. Moreover, the evaluation of richness estimators employed two complete census datasets of forest plots to be testing populations rather than using artificial datasets.