• Five in vitro gastrointestinal methods combined with SHIME (colon phase) were used.• As bioaccessibility varied in the colon phase among these methods.• Plenty of As(III) and methylated arsenicals by microbial transformation were observed.• As bioaccessibility in PBET/SBRC-SHIME colon phase were closer to in vivo results for NIST 2710a. Arsenic (As) speciation analysis is essential when evaluating the risks upon oral exposure to As-contaminated soils. In this study, we first investigated the variability in the As bioaccessibility and speciation using a combination of five common in vitro methods (SBRC, PBET, DIN, UBM and IVG) (gastric and small intestinal phases) and the SHIME model (colon phase). Our results indicate that the As bioaccessibility varies in the colon phase. An increase in the As bioaccessibility for SBRC and PBET, and a decrease for UBM and IVG were observed in the colon phase. In addition, we found different extents of methylation and large amounts of arsenite [As(III)] due to microbial reduction in the colon digests. The UBM-SHIME method displayed a higher methylation percentage of 13.5-82.1%, but a lower methylation percentage of 0.2-21.8% was observed in the SBRC-SHIME method. Besides, The MMA V levels in the colon digests were positively correlated with those of As(III) and DMA V , so DMA V can be considered an indicator to evaluate the As metabolic speed of in vitro cultured human gut microbiota. Based on the standard reference soil of NIST 2710a, the As bioaccessibility in the colon phase of PBET-SHIME and SBRC-SHIME were the closest to the in vivo results. Combining in vitro methods and SHIME will remarkably affect the accurate assessment of potential risks to human health associated with oral exposure to soil As.