Objective In a sample of 368 postmenopausal women, we (1) determined within‐cohort and between‐cohort relationships between adjuvant systemic therapy for breast cancer and self‐reported cognitive function during the first 18 months of therapy and (2) evaluated the influence of co‐occurring symptoms, neuropsychological function, and other covariates on relationships. Methods We evaluated self‐reported cognitive function, using the Patient Assessment of Own Functioning Inventory (PAOFI), and potential covariates (e.g., co‐occurring symptom scores and neuropsychological function z‐scores) in 158 women receiving aromatase inhibitor (AI) therapy alone, 104 women receiving chemotherapy followed by AI therapy, and 106 non‐cancer controls. Patients were assessed before systemic therapy and then every 6 months, for a total of four assessments over 18 months. Controls were assessed at matched time points. Mixed‐effects modeling was used to determine longitudinal relationships. Results Controlling for covariates, patients enrolled before chemotherapy reported poorer global cognitive function (p < 0.001), memory (p < 0.001), language and communication (p < 0.001), and sensorimotor function (p = 0.002) after chemotherapy. These patients reported poorer higher‐level cognitive and intellectual functions from before chemotherapy to 12 months after initiation of AI therapy (p < 0.001). Higher levels of depressive symptoms (p < 0.001), anxiety (p < 0.001), and fatigue (p = 0.040) at enrollment were predictors of poorer cognitive function over time. PAOFI total score was a predictor of executive function (p = 0.048) and visual working memory (p = 0.005) z‐scores, controlling for covariates. Conclusions Findings provide further evidence of poorer self‐reported cognitive function after chemotherapy and of relationships between co‐occurring symptoms and cognitive changes. AI therapy alone does not have an impact on self‐reported cognitive function. Copyright © 2015 John Wiley & Sons, Ltd.