The genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screen is a powerful tool used to identify therapeutic targets that can be harnessed for cancer treatment. This study aimed to assess the efficacy of genome-wide CRISPR screening to identify druggable genes associated with sorafenib-treated hepatocellular carcinoma (HCC). A genome-scale CRISPR knockout (GeCKO v2) library containing 123,411 single guide RNAs (sgRNAs) was used to identify loss-of-function mutations conferring sorafenib resistance upon HCC cells. Resistance gene screens identified SGOL1 as an indicator of prognosis of patients treated with sorafenib. Of the 19,050 genes tested, the knockout screen identified inhibition of SGOL1 expression as the most-effective genetic suppressor of sorafenib activity. Analysis of the survival of 210 patients with HCC after hepatic resection revealed that high SGOL1 expression shortened overall survival (P = 0.021). Further, matched pairs analysis of the TCGA database revealed that SGOL1 is differentially expressed. When we used a lentivirus Cas9 vector to determine the effect of targeting SGOL1 with a specific sgRNA in HCC cells, we found that SGOL1 expression was efficiently inhibited and that loss of SGOL1 was associated with sorafenib resistance. Further, loss of SGOL1 from HCC cell decreased the cytotoxicity of sorafenib in vivo. We conclude that the CRISPR screen is a powerful tool for therapeutic target analysis of sorafenib treatment and that SGOL1 serves as a druggable target for HCC treated with sorafenib and an indicator of prognosis.