Industrial agglomerations have recently gained importance in Brazil, and are being increasingly the subject of studies and public policy development. In this context, it is part of the analysis the development of typologies and classifications in order to group the agglomerations according to several characteristics such as: its history, evolution, industrial organization, amongst others. This study aims to contribute to the current discussion on industrial aglomerations by analyzing the most innovative ones. Thus, the methodologies of Cassiolato Szapiro (2003) and Mytelka and Farinelli (2000) were adjusted and applied in order to develop a ranking of the most innovative agglomerations in Brazil. For such, the data used was obtained from a research funded by the agreement SEBRAE1 / FEPESE2 / UFSC3 in 2002. Initially, 16 indicators were created and distributed in six dimensions. Also, weights were also attributed for each indicator and for each dimension - enabling greater compatibility between the methodologies applied and the one proposed by this dissertation. Finally, the grades were calculated for each agglomeration. At this moment, the methodology of Britto, Stallivieri, Campos and Vargas (2007) was of key importance in order to separate the aglomerations within groups of the same standard of innovative performance, technological learning and cooperative actions allowing comparisons between them. The aglomerations identified as the most innovative in each group were analyzed in order to identify the basic and structural characteristics, as well as the main drivers that led them into this position.