Heterologous boosting with aerosolized or intramuscular Ad5-nCoV following a two-dose CoronaVac prime has been shown to induce higher antibody levels than a homologous CoronaVac booster. However, no specific modeling has been reported to characterize the kinetics of antibody waning for these heterologous regimens. By integrating longitudinal serological data from three randomized trials conducted in Jiangsu, China (NCT04892459, NCT04952727, NCT05043259), we applied linear mixed-effects models to establish both exponential and power-law decay models for neutralizing antibodies, including live-virus neutralizing antibodies against the prototype, Delta, and Omicron BA.1 variants, and pseudovirus neutralizing antibodies against Omicron BA.4/5 variant, respectively. The findings showed that the power-law model exhibited a superior fit over the exponential model across all antibody types (all ΔAICc < 0). According to the power-law model (at day 90), the half-lives of live-virus neutralizing antibodies against the wild-type SRAS-CoV-2 strain was 195 d (95% CI: 185-210) in the aerosolized Ad5-nCoV group, 226 d (220-252) in the intramuscular Ad5-nCoV group, versus 230 d (95% CI: 222-257) in the three-dose CoronaVac group. For the Omicron BA.1 variant, the half-life was 314 d (248-453) in the aerosolized Ad5-nCoV group, 168 d (159-180) in the intramuscular Ad5-nCoV group, compared to 196 d (174-230) in the three-dose CoronaVac group. Our model indicated that the heterologous booster with Ad5-nCoV after two-dose CoronaVac, particularly the aerosolized Ad5-nCoV, induces longer-lasting neutralizing antibodies than three-dose CoronaVac, preferably characterized by power-law decay models.