The ecosystem structure and function are the foundation of ecological research. Initially, studies based on field surveys emphasized the influence of environmental gradients represented by productivity on the ecosystem structure, while later studies based on controlled experiments focused more on the biodiversity effects on ecosystem functioning such as productivity. The long-standing debates over the hypotheses and research conclusions supported by these two types of research have been the hotspot and challenge in ecology for decades. Among them, the most intriguing research field might be whether a general relationship between biodiversity and productivity exists in natural ecosystems or not. With a thorough bibliometric analysis and literature review of the biodiversity-productivity relationship research based on the Web of Science Core Collection, this study aims to explore the research history of the relationship between biodiversity and productivity, understand how it promotes the theories and paradigm shift of the ecological research, thus raising suggestions for future directions. Our results show that: (1) the biodiversity-productivity relationship study promotes the paradigm shift from observational research based on filed surveys to experimental studies based on controlled experiments, and inspires the development of global research networks such as the Nutrient Network (NutNet); (2) the studied ecosystems have expanded from prairie in North America to diverse ecosystems such as other grasslands (e.g. savanna and Eurasian Steppe), shrubs and forests, which enhances the universality of the conclusions and underlying mechanisms of the biodiversity-productivity relationship; (3) these studies promotes the understanding of the role of different biodiversity facets in maintaining ecosystem functions, e.g. functional and phylogenetic diversity, and stimulates the research on the ecosystem multifunctionality beyond the single function such as productivity; and (4) there are large differences in experimental design and theoretical framework among resource groups, particularly between the Northern Europe and the United States, leading to a fragmented research network which impedes the improvement and integration of different ecological hypotheses and probably contributes to the longstanding debates over the relationship between biodiversity and ecosystem functions. Future ecological research should pay more attention to the boundaries and controlling factors of different hypotheses, thus allowing for an integration of different hypotheses with a single unified framework to understand the general relationship between ecosystem structure and function in natural ecosystems.