To meet the demand for test scenarios in autonomous driving simulation testing, this study proposed a method for generating a simulation test scenario library that considers the static characteristics of cut-in vehicles. This research focused on highway lane-change cut-in scenarios, developing a logical scenario model that integrates motion information and static characteristics of cut-in vehicles, based on lane-change cut-in data from the highD dataset. A multi-class Gaussian mixture model constructed the joint probability density function of the logical scenario model parameters. The adaptive importance sampling (AIS) algorithm was utilized as the sampling method, and a time-to-collision (TTC) weighting function was introduced to formulate the scenario library generation method, which enabled the generation of a simulation test scenario library according to specific requirements. Experimental results indicate that, in comparison to traditional Markov Chain Monte Carlo and importance sampling methods, the AIS-based scenario library generation method achieves a reduction in the jensen-shannon divergence between the scenario parameter distribution and the original distribution by 65.24% and 79.72%, respectively, and results in a reduction in the number of effective test scenarios by 96.22% and 90.37%, respectively. The proposed method ensures comprehensive coverage of scenario generation while enhancing testing efficiency. Moreover, by adjusting the TTC weighting function, the method increases the proportion of high-risk test scenarios, thereby providing an adequate number of high-risk scenarios for autonomous driving testing, fulfilling diverse testing requirements.