Objective: Mutant KRAS, the principal isoform of RAS, plays a pivotal role in the oncogenesis of colorectal cancer by constitutively activating the RAF/MEK/ERK and PI3K/AKT pathways. Effective targeted therapies are urgently needed. We investigated whether rigosertib, a benzyl styryl sulfone RAS signaling disruptor, could selectively kill KRAS-mutant colorectal cancer cells.Methods: CCK-8 was used to determine the cell viability. Patient-derived tumor and cancer cell xenograft models were used to detect the inhibitory efficacy of rigosertib. Flow cytometry was used to evaluate the apoptosis and cell cycle progression. Apoptosis and cell cycle arrest markers were detected by Western blot. DCFH-DA was used to determine the reactive oxygen species. Immunohistochemistry staining and Western blot were performed to characterize RAS signaling markers in colorectal cancer tissues and cells. Results: Rigosertib (RGS) exhibited a cytotoxic effect against colorectal cancer cells, which was greater in KRAS-mutant cells. Furthermore, RGS induced mitotic arrest and oxidative stress-dependent apoptosis in KRAS-mutant DLD1 and HCT116 cells. Besides, RGS disrupted RAS signaling, and the inhibition of RAS/MEK/ERK was independent of cellular oxidative stress. Using patient-derived xenograft models, the response and tumor inhibition of RGS were significantly higher in the KRAS-mutant subgroup, while p-MEK, p-ERK, and p-AKT levels of RGS-treated tumors were significantly decreased. Finally, in a KRAS-mutant, chemotherapy-resistant patient-derived xenograft model, RGS showed a stronger therapeutic effect than the combination standard therapy involving fluoropyrimidine + oxaliplatin/irinotecan + bevacizumab. Conclusions: These data showed that targeting RAS signaling using RGS could be a therapeutic treatment for KRAS-mutant colorectal cancer patients.