The optimal dose intensity of chemotherapy for elderly patients with diffuse large B cell lymphoma (DLBCL) remains controversial because of concerns about adverse events and comorbidities related to the patients' frailty. This single-center study retrospectively analyzed patients aged ≥ 70 years who were newly diagnosed with DLBCL and received chemotherapy between 2004 and 2022. Survival outcomes and treatment-related mortality (TRM) were stratified according to geriatric assessment variables, and the influence of chemotherapy dose intensity on outcomes was assessed using the frailty score with a Cox hazards model with restricted cubic spline (RCS) in patients aged 70-79 years. In total, 337 patients were included. The frailty score accurately predicted prognosis (5-year overall survival [OS]: 73.1%, 60.2%, and 29.7% in fit, unfit, and frail patients, respectively; P < 0.001) and TRM (5-year TRM: 0%, 5.4%, and 16.8 in fit, unfit, and frail patients, respectively; P < 0.001). Cox regression with RCS demonstrated a linear association between dose intensity and survival outcomes. Initial dose intensity (IDI) and relative dose intensity (RDI) had a significant impact on OS in fit patients. However, IDI and RDI had no significant effect on survival in non-fit (unfit and frail) patients. The frailty score identified non-fit patients with poorer survival and a higher risk of TRM. While fit patients were likely to benefit from full-dose R-CHOP, unfit and frail patients would likely benefit more from attenuated R-CHOP. This study suggested a potential role for the frailty score in individualizing treatment intensity in elderly patients with DLBCL.