Mutation of genes related to the SWI/SNF chromatin remodeling complex is detected in 20% of all cancers. The SWI/SNF chromatin remodeling complex comprises about 15 subunits and is classified into three subcomplexes: cBAF, PBAF, and ncBAF. Previously, we showed that ovarian clear cell carcinoma cells deficient in ARID1A, a subunit of the cBAF complex, are synthetic lethal with several genes required for glutathione (GSH) synthesis and are therefore sensitive to the GSH inhibitor eprenetapopt (APR-246). However, we do not know whether cancer cells deficient in SWI/SNF components other than ARID1A are selectively sensitive to treatment with eprenetapopt. Here, we show that SMARCA4-, SMARCB1-, and PBRM1-deficient cells are more sensitive to eprenetapopt than SWI/SNF-proficient cells. We found that deficiency of SMARCA4, SMARCB1, or PBRM1 attenuates transcription of the SLC7A11 gene (which supplies cysteine as a raw metabolic material for GSH synthesis) by the failure of recruitment of cBAF and PBAF to the promotor and enhancer regions of the SLC7A11 locus, thereby reducing basal levels of GSH. In addition, eprenetapopt decreased the amount of intracellular GSH and increased the intracellular amount of reactive oxygen species (ROS), followed by induction of apoptosis. Taken together, eprenetapopt could be a promising selective agent for SWI/SNF-deficient cancer cells derived from SMARCA4-deficient lung cancers, SMARCB1-deficient rhabdoid tumors, and PBRM1-deficient kidney cancers.