In the context of climate change, drought events were likely to be frequent and severe in Fenhe river basin (FRB). This study developed a multivariate composite drought index (CDI) combining standardized precipitation index (SPI), standardized runoff index (SRI), and agricultural standardized precipitation index (aSPI) based on entropy weight method for comprehensive characterizing drought in the FRB. A mixed copula was then developed for comprehensively modeling the dependent structures of drought variables (duration, severity, peak), through coupling three single copulas (Gumbel, Clayton and Frank copulas). The drought variation was examined by modified Mann-Kendall test (MMK) method and the historical droughts of the basin were examined in three periods: the first period (1951–1974), the second period (1975–1998), and the third period (1999–2022). The bivariate return periods and bivariate risk of drought variables were finally analyzed based on mixed copula. The results indicate that (1) the constructed CDI is consistent with SPI, SRI and aSPI series, and it is sensitive and effective to capture historical drought events. (2) the CDI, SPI, SRI and aSPI present significant decreasing trend indicating the increasingly serious drought condition in the basin. (3) drought variables are higher in the third period than in other periods and the basin experienced an increasingly serious dry condition. (4) the mixed copula provides more reliable return periods and drought risk results.