Dissolved organic matter (DOM) is a major carbon reservoir in aquatic environments. It is composed of hundreds of thousands of molecular species that play a crucial role in diverse ecosystems. While the molecular diversity is influenced by the activity of heterotrophic microbes, the impact of substrate concentration on the resulting DOM composition remains poorly understood. A clear understanding of this relationship is critical for anticipating how the ecological roles of DOM may be altered by environmental changes such as eutrophication. This study aims to elucidate how changes in substrate concentration affect the composition of the microbiome and metabolites, and to examine how these changes influence the molecular diversity of DOM. To begin the experiment, surface water from the northern basin of Lake Biwa (Japan) was filtered (1.2 μm GF/C) and diluted 50-fold by ultrapure water to prepare an artificial lake water medium in 2 L PC bottles. Glucose was added at concentrations of 3, 6, and 12 mg C L-1, each in duplicate. Nutrients and trace metals were supplemented to match natural concentrations in Lake Biwa. Bottles were incubated in the dark at 20°C while shaking at 60 rpm. Samples were collected on days 24, 49, 76, 207, 410, and 613, then filtered (0.1 μm), acidified to pH 2, and extracted using PPL resin. The extracted DOM was analyzed by ultrahigh-resolution mass spectrometry (7T solariX, Bruker), identifying 5,232 unique molecular formulas. Changes in DOM composition were assessed using principal component analysis (PCA) after robust centered logratio transformation and the effective number of species. DNA was extracted from filter residues, and 16S rRNA amplicon sequencing was performed. After removing singletons, 521 amplicon sequence variants (ASVs) were identified. Microbial community composition was analyzed using the same statistical framework applied to DOM data based on ASV relative abundance. In all cultures, DOC concentration rapidly decreased at the beginning due to glucose consumption, reaching 0.22 to 0.46 mg C L-1 by day 24, and remained stable afterwards. PCA revealed temporal changes in DOM molecular composition. Also, the microbiome changed over time depending on the initial glucose concentration. Regarding species diversity, the molecular diversity of DOM in the stationary phase tended to increase with higher initial glucose concentrations. In contrast, microbial diversity decreased under higher initial glucose due to reduced species evenness. Our results show that in closed systems, higher substrate levels increase the molecular diversity of microbial DOM, with potential impacts on ecosystem functioning.