INTRODUCTION: Older adults with cancer have unique fall risk factors related to their disease and treatment such as polypharmacy and neurotoxic treatments. In this secondary analysis, we identified modifiable risk factors associated with future falls among older adults with advanced cancers. MATERIALS AND METHODS: Data were from the COACH study (ClinicalTrials.gov: NCT02107443; PI: Mohile). Patients were age  ≥  70, had stage III/IV solid tumor or lymphoma, ≥ 1 geriatric assessment impairment, and were receiving palliative intent treatment. Falls were self-reported at baseline (in the past six months), four to six weeks, three months, and six months. We generated inverse probability weights to account for mortality-related loss to follow-up and applied these in generalized linear mixed models to estimate incidence rate ratios. RESULTS: Of 541 patients (mean age: 77, standard deviation [SD]: 5.27), 140 (26%) reported prior falls at baseline, and 467 (86%) had falls data for ≥ 1 follow-up timepoint. Of those, 103 (22%) reported at least one fall during the follow-up period, and 112 (24%) had incomplete follow-up due to death. In fully adjusted models, prior falls and impaired Timed Up and Go score were associated with higher incidence of falls over 6 months. DISCUSSION: We identified several potentially modifiable fall risk factors in older adults with advanced cancers. Future studies should consider ways to integrate fall risk assessment into ongoing cancer care and intervene to reduce falls in this population.