This study aimed to investigate the interaction between methane production performance and active microbial community dynamics at different loading rates by increasing influent substrate concentration. The model system was an upflow anaerobic sludge blanket (UASB) reactor using molasses wastewater. The active microbial community was analyzed using a ribosomal RNA-based approach in order to reflect active members in the UASB system. The methane production rate (MPR) increased with an increase in organic loading rate (OLR) from 3.6 to 5.5 g COD center dot L-1 center dot day(-1) and then it decreased with further OLR addition until 9.7 g COD center dot L-1 center dot day(-1). The UASB reactor achieved a maximum methane production rate of 0.48 L center dot L-1 center dot day(-1) with a chemical oxygen demand (COD) removal efficiency of 91.2 % at an influent molasses concentration of 16 g COD center dot L-1 (OLR of 5.5 g COD center dot L-1 center dot day(-1)). In the archaeal community, Methanosarcina was predominant irrespective of loading rate, and the relative abundance of Methanosaeta increased with loading rate. In the bacterial community, Firmicutes and Eubacteriaceae were relatively abundant in the loading conditions tested. The network analysis between operation parameters and microbial community indicated that MPR was positively associated with most methanogenic archaea, including the relatively abundant Methanosarcina and Methanosaeta, except Methanofollis. The most abundant Methanosarcina was negatively associated with Bifidobacterium and Methanosaeta, whereas Methanosaeta was positively associated with Bifidobacterium.