Abstract: Interval-censored data is a common data type in survival analysis, and the failure time we are interested in is dependent on the observation time in practical problems. Ignoring the dependence may lead to biased analysis results. We consider the semi-parametric regression analysis of additive-multiplicative hazards models with informative Case Ⅱ interval-censored data. We introduce a frailty to characterize the correlation among the observation time and the failure time, and construct the corresponding additive-multiplicative frailty model for the failure time and the observation time, respectively. In order to estimate the parameters, we propose a Sieve maximum likelihood estimation method based on Bernstein polynomials, and combine it with EM algorithm to estimate the parameters of interest. Then, an extensive simulation study suggests that the proposed method performs well. Finally, the proposed method is applied to the real data generated in the age-related eye disease study, and the analysis revealed that the model exhibited excellent fitting performance.