Exercise may be a strategy for improvement of cognitive deficits commonly present in people with idiopathic generalized epilepsies (IGE). We investigated the relationship between cognition and level of physical exercise in leisure (PEL) in people with IGE who have been seizurefree for at least 6 months (IGE−) as compared to those who have not been seizurefree (IGE+) and healthy controls (HCs). We hypothesized that higher level of physical exercise is associated with better cognitive functioning in patients with IGE and HCs, and that seizure control affects both PEL levels and cognitive functioning in patients with IGE. We recruited 75 participants aged 18–65: 31 people with IGE (17 IGE−, 14 IGE+) and 44 HCs. Participants completed assessments of quality of life (SF-36), physical activity levels (Baecke questionnaire and International Physical Activity Questionnaire (IPAQ)) and cognition (Montreal Cognitive Assessment (MoCA), Hopkins Verbal Learning Test – Revised (HVLT), and flanker task). Group differences (HCs vs. IGE; HCs vs. IGE+ vs. IGE−) were assessed. Pearson correlations examined linear relationships between PEL and cognitive performance. Groups were similar in age and sex. Compared to HCs, patients with IGE had higher body mass index, fewer years of education, and consistently scored worse on all measures except flanker task accuracy on incongruent trials. When examining IGE− and IGE+ subgroups, compared to HCs, both had higher body mass index, and fewer years of education. Healthy controls scored significantly better than one or both of the IGE groups on SF-36 scores, PEL levels, IPAQ activity level, MoCA scores, HVLT learning and long-delay free-recall scores, and flanker task accuracy on congruent trials. Among patients with IGE, there were no significant differences between age of epilepsy onset, duration of epilepsy, number of anti-seizure drugs (ASDs) currently being used, or the group distribution of type of IGE. In the combined sample (IGE+, IGE− and HCs), PEL positively correlated with MoCA scores (Pearson's r = 0.238; p = 0.0397) and with flanker task accuracy on congruent trials (Pearson's r = 0.295; p = 0.0132). Overall, patients with IGE performed worse than HCs on cognitive and physical activity measures, but the cognitive impairments were more pronounced for IGE+, while physical exercise levels were less for patients with IGE regardless of seizure control. While positive relationships between leisure-time PEL and cognitive performance are promising, further investigations into how exercise levels interact with cognitive functioning in epilepsy are needed.