Endometriosis is a potential premalignant disorder. The underlying molecular aberrations, however, are not fully understood. A recent exome sequencing study found that 25% (10/39) of deep infiltrating endometriosis harbored cancer driver gene mutations. However, it is unclear whether these mutations also exist in ovarian endometriosis. Here, a total of 101 ovarian endometriosis samples were analyzed for the presence of these gene mutations, including KRAS, PPP2R1A, PIK3CA and ARID1A. In addition, 6 other cancer-associated genes (BRAF, NRAS, HRAS, ERK1, ERK2 and PTEN) were also analyzed. In total, four somatic mutations were identified in three out of 101 ovarian endometriotic lesions (4%, 4/101), including a KRAS p.G12V, a PPP2R1A p.S256F and two ARID1A nonsense mutations (p.Q403* and p.G1926*); while no mutations were identified in the remaining 7 genes (BRAF, NRAS, HRAS, ERK1, ERK2, PTEN and PIK3CA). Note that the KRAS G12V and ARID1A Q403* mutations co-occurred in a 36-year-old sample who had a high serum CA125 (308.4 U/mL) and a late menarche age (18-year-old). Additionally, no mutations in any of the 10 genes were identified in either the healthy eutopic endometrial tissues from 85 control individuals without endometriosis, or in 62 healthy ovarian tissues from ovarian cysts samples (without endometriosis). Our study revealed, for the first time, the presence of classical cancer driver gene mutations in ovarian endometriosis. Furthermore, the co-occurrence of KRAS and ARID1A mutations was identified in a single individual for the first time. The observations of cancer driver gene mutations in our ovarian endometriosis samples, together with several prior observations, further support the notion that endometriosis is a premalignant disorder.