Background: Nutritional status is an important factor influencing toxicity of treatment. Nutritional assessment indicators such as the Prognostic Nutritional Index (PNI), Controlling Nutritional Status (CONUT) score and modified Glasgow Prognostic Score (mGPS) have been reported to be associated with treatment-related adverse events (AEs) for various malignancies. However, there are no reports investigating the relationship between nutritional status and AEs from poly-(ADP-ribose) polymerase (PARP) inhibitors (PARPi), which are widely used in recent years as maintenance therapy for ovarian cancer. Objective: The primary objective was to investigate the usefulness of nutritional assessment indicators in predicting treatment discontinuation due to AEs from PARPi. Methods: This multicenter retrospective study included patients diagnosed with ovarian cancer who received maintenance therapy with PARPi from January 2018 to December 2023. PNI, CONUT score, and mGPS were calculated based on hematological parameters measured within 7 days before the start of PARPi therapy. Results: A total of 272 patients received maintenance therapy with PARPi during the period, but due to the absence of the blood collection of albumin levels within one week or other exclusion criteria, 71 patients were finally included in this analysis. AEs were seen in 59 patients (83.1%), including 25 (35.2%) severe events (grade ≥3 in Common Terminology Criteria for Adverse Events v5.0). Eighteen patients (25.4%) discontinued treatment due to PARPi-related AEs. Low PNI (<48.44) and high mGPS (≥1) were predictors of treatment discontinuation in both univariate and multivariate analyses. CONUT was not a significant predictor in this study. Conclusions: Our study suggested that PNI and mGPS can predict the risk of treatment discontinuation due to PARPi-related AEs before starting maintenance therapy. This insight opens avenues for more personalized treatment plans, potentially improving patient outcomes.