Objective Subjective cognitive complaints are a frequent concern of patients with epilepsy. The Aldenkamp–Baker Neuropsychological Assessment Schedule (ABNAS) is a patient-reported scale validated to measure adverse cognitive effects of antiepileptic drugs (AEDs). The goals of this study were to identify predictors of patient-reported cognitive dysfunction and to assess the relationship between subjective and objective cognitive impairment. Methods The Cleveland Clinic Knowledge Program Data Registry was used to identify adult patients seen in outpatient epilepsy clinic from January to May 2015 and who completed the following scales: ABNAS for subjective cognitive impairment, Patient Health Questionnaire (PHQ-9) for depression, Generalized Anxiety Disorder 7-item (GAD-7) scale, Quality of Life in Epilepsy (QOLIE-10), and EuroQOL five dimensions questionnaire (EQ-5D) for health-related quality of life. Topiramate (TPM) was considered a high-risk medication for cognitive impairment. Patients were categorized into groups based on total ABNAS score: subjective cognitive impairment (ABNAS > 15; N = 270) and no subjective cognitive impairment (ABNAS ≤ 15; N = 400). Multivariable logistic regression models were constructed to identify independent predictors of subjective cognitive impairment. In a subset of patients who had neuropsychological testing within 6 months of completing the ABNAS (N = 60), Pearson correlations and multivariable logistic regression models, controlling for number of AEDs, depression, and anxiety, assessed the relationship between subjective cognitive impairment and objective cognitive performance on measures of intelligence, attention/working memory, verbal fluency, naming, processing speed, manual dexterity, visuomotor processing, and verbal memory. Results Forty percent of patients in the overall sample (N = 270/670) reported cognitive impairment. The variables most strongly associated with subjective cognitive impairment were PHQ-9 score, number of AEDs, and seizure frequency. In the subset of patients with neuropsychological testing, ABNAS score was correlated with anxiety (r = 0.44), depression (r = 0.38), and attention/working memory (r = − 0.31). After adjusting for depression and anxiety, patients who endorsed subjective cognitive impairment scored significantly lower on measures of nonverbal intelligence and attention/working memory, but not on other cognitive measures. Conclusions Subjective cognitive impairment as reported on the ABNAS is most strongly associated with depressive symptomatology, number of AEDs, and seizure frequency, but not with most objective cognitive measures. Identifying these three predictors provides a clear framework to understand and address subjective cognitive complaints in adult patients with epilepsy. [ABSTRACT FROM AUTHOR]