This paper employs a Multiscale Geographically Weighted Regression (MGWR) model to examine the spatial variation of regional income convergence across Indonesia. This new MGWR model provides a multiscale analysis of spatial heterogeneity. Given the large socioeconomic, demographic, and geographic differences across Indonesian islands, models based on spatial heterogeneity are likely to provide a more accurate and realistic assessment of a regional convergence process. Most previous studies have largely ignored the role of spatial heterogeneity. They usually report a common convergence speed for all subnational regions. In contrast, this paper departs from the assumption of a common convergence speed and aims to identify multiple spatial clusters that characterize the heterogeneity of the regional convergence process of Indonesia.