In Brazil, there is a concern with the quality of education, especially when analyzing the results of large-scale evaluations both at the national level, by the Basic Education Development Index (Ideb), and at the international level, by the Program for International Student Assessment (PISA). Thus, as a way of contributing to this issue, this study aimed to identify the best practices that can help improve the performance of Brazilian municipal elementary schools, considered of low socioeconomic level (NSE) in Ideb. The method used was quali-quantitative, where mathematical models were used, such as Data Envelopment Analysis (DEA), Quintile Analysis, Logistic Regression and case studies. As a result, 14 best practices were identified that may have contributed to the performance of schools considered efficient. It is hoped that the study can contribute to the enrichment of research in the area, in addition to the financial investment decision process, allocation of public resources and educational policies, through an efficient school management that aims to improve the quality of education in Brazil.