The complementary relationship (CR) of evapotranspiration is of utmost importance for evapotranspiration estimation using only routine meteorological data in humid lowland areas in Southeast China, which are characterized by subtropical monsoons and dense river networks. This study evaluated the performance of five CR-based evapotranspiration models, namely, the advection‐aridity (AA), Granger and Gray (GG), B2015, H2018, and C2018 models, and the asymmetry in the CR between the actual evapotranspiration (ET_a) and apparent potential evapotranspiration (ET_pa), based on four years of measured data. The results indicated that the C2018(T_W1) model with a simple structure is the best choice (among the models considered) when no calibration is feasible. After calibration, most of the models showed an excellent agreement with observations with NSE ≥ 0.89 and R²≥ 0.91 at the daily and monthly timescales. The H2018 model achieved a superior performance to that of the other models at the daily and monthly scales. A significantly asymmetric CR was observed between daily ET_a and ET_pa values, and scatter points were clustered within the 0.7–1 humidity range (ET_a/ET_pa).