Different dog sizes are associated with variations in large intestinal physiology including gut microbiota, which plays a key role in animal health. This study aims to evaluate, using the CANIM-ARCOL (Canine Mucosal Artificial Colon), the relative importance of gut microbes versus physicochemical and nutritional parameters of the canine colonic environment in shaping microbiota structure and functions. CANIM-ARCOL was set up to reproduce nutrient availability, bile acid profiles, colonic pH, and transit time from small, medium, or large dogs according to in vivo data, while bioreactors were all inoculated with a fecal sample collected from medium size dogs (n = 2). Applying different dog size parameters resulted in a positive association between size and gas or SCFA production, as well as distinct microbiota profiles as revealed by 16S Metabarcoding. Comparisons with in vivo data from canine stools and previous in vitro results obtained when CANIM-ARCOL was inoculated with fecal samples from three dog sizes revealed that environmental colonic parameters were sufficient to drive microbiota functions. However, size-related fecal microbes were necessary to accurately reproduce in vitro the colonic ecosystem of small, medium, and large dogs. For the first time, this study provides mechanistic insights on which parameters from colonic ecosystem mainly drive canine microbiota in relation to dog size. The CANIM-ARCOL can be used as a relevant in vitro platform to unravel interactions between food or pharma compounds and canine colonic microbiota, under different dog size conditions. The potential of the model will be extended soon to diseased situations (e.g. chronic enteropathies or obesity).