Mann-Kendall mutation diagnosis method and serial cluster analysis method were employed to diagnose the variation points of runoff series in the Beiluo River Basin during dry-flood alternating period, flood period, flood-dry alternating period, and dry season, and the variation grades were classified. The runoff reduction method based on SIMHYD model was applied to calculate and distinguish the contribution rates of climate change and human activities to runoff variation in different periods. The results show that the runoff variation grade in different periods ranges from grade zero to grade four. Excluding the non-variation point of grade 0, the variation occurred in all periods. The runoff variation grade is related to runoff. The runoff variation grade is lower in the period of large runoff(flood season, dry-flood alternating period)and higher in the period of small runoff(flood-dry alternating period, dry season). The lowest dry period of runoff is 4-level strong variation. The contribution rate of human activities to the variation of runoff in the dry-flood alternating period is larger, and human activities are the main driving force. The contribution rate of climate change in the remaining three periods is larger, and climate change is the main driving force.