Aim The present study aimed to investigate whether preoperative medication use is associated with postoperative length of hospital stay in older adults undergoing cancer surgery. Methods Patients aged ≥65 years who were scheduled for cancer surgery and presented for preoperative comprehensive geriatric assessment were included in the present study. Cognitive function evaluation and preoperative medication review were carried out, as well as baseline characteristics of participants collected from electronic medical records. The primary efficacy variable was the postoperative length of stay (LOS) in hospital. Results A total of 475 cancer patients were included for the analysis. Baseline characteristics of participants including older age, lower body mass index (BMI) and male sex were associated with longer postoperative stay. Among the clinical variables, cancer type, number of medications, potentially inappropriate medication (PIM) and delirium‐inducing medication were found as statistically significant factors for postoperative LOS. In multivariate analysis, variables independently associated with postoperative LOS were cancer type, PIM use, BMI, and the number of medications after controlling for age, BMI, sex, cancer type, the number of medications, PIM, and delirium‐inducing medication. In subgroup analysis of gastrointestinal cancer, multiple linear regression analysis showed that PIM use and BMI were significantly associated with LOS after adjustment for age, sex, and number of medication. Conclusions The present study supports the impact of medication use on postoperative LOS in geriatric oncology patients. The results add a further aspect to medication optimization in older patients undergoing cancer surgery. Geriatr Gerontol Int 2017; ••: ••–••.