This article discusses collaborative models of social indexing and their applicability in digital libraries. The objective is to identify in the scientific literature of information science the major studies whose focus is related to the construction of models that propose the integration of folksonomy in metadata for information retrieval systems. It discusses aspects of the folksonomy from the perspective of document description processes in a digital environment, identifying types of collaborative models whose rules can be improved or directly applied to the collaborative indexing of informational objects in digital libraries. The research methodology is bibliographical and exploratory, based on literature review on folksonomy, collaborative models of social indexing, metadata and digital libraries. It concludes that there is a need to create methodologies for use of user-generated metadata, aiming to improve information representation in digital libraries.