Autism spectrum disorders (ASD) make up a heterogeneous group of neurodevelopmental disorders characterized by social and communication difficulties associated with repetitive and restrictive behaviors. Besides core features, metabolic imbalances, inflammation, gastrointestinal (GI) symptoms, and altered gut microbiota composition were often described in association with ASD, but their connection with the severity of autism (SA) remains unexplored. In this study, fecal metabolome, microbiota, and calprotectin levels of 80 ASD preschoolers were quantified and correlated with SA. Twelve of the fifty-nine molecules that were quantified by fecal metabolome analysis were significantly associated with SA. No links between SA or GI symptoms and microorganisms’ relative abundance were highlighted. Significant correlations between bifidobacteria, Sutterella, lactobacilli relative abundance, and metabolomics profiles were found. These results suggest that fecal metabolome discriminates the SA and intestinal microorganisms mediate the link between metabolome and SA regardless of GI symptomatology. The study raises the possibility that grouping ASD populations through metabolomics and fecal microbiota could aid the identification of specific ASD endophenotypes, on the basis of the SA. Mechanistic studies focusing on detected biomarkers might be an option for future studies.