In the context of doping control, conventional direct chemical testing detects only a limited scope of target substances in equine biological samples. To expand the ability to detect doping agents and their detection windows, metabolomics has recently become a common approach for monitoring alteration of biomarkers caused by doping agents in relevant metabolic pathways. In horse racing, remarkable changes in metabolic profiles between the rest state and racing are likely to affect the identification of doping biomarkers. Previously, we reported a limited number of significantly upregulated metabolites after racing, based on a non‐targeted metabolomics approach using out‐of‐competition and post‐race equine plasma samples. In this study, we performed a more thorough analysis of the data set, using pathway analysis to establish a post‐race biomarkers database (PBD) that includes upregulated and downregulated metabolites, their fold changes, and relevant pathways, with the main objective of improving our understanding of changes in physiological status related to horse racing. Statistical analysis of the PBD revealed that two peak combinations of pentadecanoyl carnitine/galactosylglycerol (P/G) and heptadecanoyl carnitine/galactosylglycerol (H/G) could be used as potential biomarkers for the discrimination of the rest and post‐race groups. To demonstrate the applicability of the PBD, we validated the post‐race biomarkers P/G and H/G (highly involved in lipid metabolism) by a single‐blind test. This strategy, which combines establishment of a biomarker database with pathway analysis, represents a powerful tool for discovering potential doping biomarkers in the future.