Epilepsy is one of the most common neurological diseases and it causes profound morbidity and mortality. We identified the first de novo variant in KCNMA1 (c.2984 A > G (p.(N995S)))-encoding the BK channel-that causes epilepsy, but not paroxysmal dyskinesia, in two independent families. The c.2984 A > G (p.(N995S)) variant markedly increased the macroscopic potassium current by increasing both the channel open probability and channel open dwell time. The c.2984 A > G (p.(N995S)) variant did not affect the calcium sensitivity of the channel. We also identified three other variants of unknown significance (c.1554 G > T (p.(K518N)), c.1967A > C (p.(E656A)), and c.3476 A > G (p.(N1159S))) in three separate patients with divergent epileptic phenotypes. However, these variants did not affect the BK potassium current, and are therefore unlikely to be disease-causing. These results demonstrate that BK channel variants can cause epilepsy without paroxysmal dyskinesia. The underlying molecular mechanism can be increased activation of the BK channel by increased sensitivity to the voltage-dependent activation without affecting the sensitivity to the calcium-dependent activation. Our data suggest that the BK channel may represent a drug target for the treatment of epilepsy. Our data highlight the importance of functional electrophysiological studies of BK channel variants in distinguishing whether a genomic variant of unknown significance is a disease-causing variant or a benign variant.