<p>The Paris Agreement, aimed at limiting the increase in global temperatures and reducing greenhouse gas emissions, has encouraged many countries, including Brazil, to promote the production and use of biofuels as part of their strategies to mitigate climate change. In this context, cotton cultivation, particularly in the state of Mato Grosso, has increased, positioning the country as the fourth-largest producer globally and the second-largest exporter. However, this expansion necessitates the exploration of alternatives that enhance the value of the waste generated in the cotton production chain, contributing to agricultural efficiency, the circular economy, and, consequently, sustainable development. In this research, applied unsupervised computational methods, specifically natural language processing, structural topic modeling and complex network analysis, to map knowledge domains related to cotton biorefineries. The adopted approach is innovative and enables the mapping of the current state of technologies and trends in the use of specific products. Basic information, such as the identification of inputs generated throughout the production process, served as a foundation for article research. This procedure facilitated the formation of citation networks and the identification of new technologies and trends. The results obtained from the analysis of the selected articles allowed for the development of a layout for a cotton biorefinery, integrating processes that add value to the waste, with identified technologies ranging from the production of bioplastics, substrates, and resins to animal feed.</p> Graphical Abstract <p></p>