Abstract: Traffic variation tendency has a significant impact on the efficiency and safety of urban traffic operations. To explore and improve its traffic efficiency in adjacent weaving segments (AWS) near a ramp of an urban rapid road, this paper proposes an adaptive traffic state discrimination method combining the cusp-mutation theory and Gaussian-mixture model. With an aim of reducing traffic congestion, the optimal combination matching of the cusp-mutation theory and Gaussian-mixture model is discussed to establish a more practical determining model for the actual traffic operation status of the AWS in this paper. This method focuses on the AWS of urban rapid roads, takes the AWS as the research object, and studies the Gaussian-mixture distribution of the average speed and time occupancy under different spacing conditions. At the same time, it takes the advantage of the cusp-mutation theory to analyze the trend of traffic variation in the AWS and uses the Gaussian-mixture model to identify and classify the congestion distribution pattern of the traffic operation status in the AWS. Determining traffic is the theoretical basis, which provides necessary theoretical and technical support for traffic control in the AWS by establishing the traffic flow model and the state discrimination model of AWS. Besides the integration of the cusp-mutation theory and the Gaussian-mixture model, a method for estimation of traffic state parameters is proposed to solve the problem that the congestion state discrimination criteria of the AWS are affected by the weaving interval distance. Through investigating the actual road structure and traffic volume, the real-time traffic operation status and its changes under the condition of an increasing traffic flow are simulated and analyzed using the VISSIM traffic simulation software for the selected AWS at the interchange of Haixia Road and Sigongli Road in Chongqing. It is verified that the cusp-mutation theory is applicable to discriminating operating status of the AWS, and the critical state of traffic is a key to evaluating level of traffic congestion, which changes along with different distances between the two weaving segments. The experimental results show that the critical congestion velocity varies significantly when the spacing between two AWS is less than 150 m, and stabilizes when it is greater than 150 m. The critical congestion rate varies significantly when the spacing between two AWS is less than 200 m, and stabilizes when the spacing is greater than 200 m. This study highlights the importance of an adaptive traffic state discrimination method that combines the cusp-mutation theory and the Gaussian-mixture model to study the traffic flow in the AWS. The effectiveness of the proposed method in analyzing traffic state change is also demonstrated experimentally, which helps to study traffic control and traffic flow in the AWS, provides a strong theoretical basis for analyzing the trend of traffic state change in the AWS, and offers valuable insights into the effect of different spacing conditions on the critical congestion speed and rate. The study can help traffic managers to formulate better traffic control strategies and improve the traffic efficiency of the AWS.