Abstract The study of metabolomics requires extracting as many metabolites as possible from a biological sample. This study aimed to determine the optimal method for the extraction of metabolites from solid-state fermented cottonseed meal (FCSM). The UPLC-Q-TOF-MS global metabolomics technology was used to detect the metabolites in FCSM, and the extraction quantity and extraction efficiency of seven different extraction methods, specifically the WA, 50MeOH, 50MeOHB, 50MeCNB, 80MeOHB, 80MeOH and AMF methods were evaluated. The results showed that the number of VIP metabolites extracted by AMF method are 196 and 184 in ESI+ and ESI- mode respectively, it is the largest number of all exacted methods; and the AMF methods also provided a higher extraction efficiency compared with the other methods, especially in indoleacrylic acid, dl-tryptophan and epicatechin (p < 0.01). As a result, AMF/-4 °C method was identified as the best method for the extraction of metabolites from FCSM by Lactobacillus acidophilus. Our study establishes a technical basis for future metabolomics research of fermented feed.