Background: In temporal lobe epilepsy, seizures tend to occur at preferential times of the day. This timing of seizures may result from the sleep-wake history (process S) and/or from the circadian rhythm (process C). We aimed at disentangling their contributions to timing limbic seizures in the hippocampal kainic acid mouse model. Methods: Using 16 depth-electrodes, we recorded limbic seizures in epileptic mice (n=24) exposed to different chronobiological paradigms over months: 1)12h:12h light-dark cycle (LD), 2) gentle sleep deprivation (GSD), 3) continuous dim red light (DD), 4) continuous light (LL), 5) 10h:10h light-dark cycle (T20). We tested the statistical association of seizures with the physiological circadian cycle (measured as core temperature) and with the sleep-wake state and history, automatically annotated in month-long EEG recordings using non- negative matrix factorization. Results: In LD seizures occurred during sleep and wakefulness and showed mild to moderate circadian clustering after the circadian peak (PLV= 0.25+/-0.13). This circadian clustering persisted with similar phase and strength in DD (PLV= 0.33+/-0.11) but was attenuated in LL (PLV= 0.26+/-0.03), when mice slept longer. Forced desynchrony of the circadian and sleep-wake processes was obtained in the T20 paradigm, and circadian seizure clustering was stronger when the two processes aligned (PLV= 0.41+/-0.11) and weaker when misaligned (PLV = 0.18+/-0.7). GSD led to an increase in seizure rate coinciding with an increase in slow wave activity. Conclusion: Circadian seizure timing depends on the strength and alignment of Process S and Process C. This finding may guide future chronotherapeutic interventions.