In order to improve the timing of the signal lights of the single intersection and adjacent intersections and improve the traffic efficiency, the optimal design of the single intersection cannot fully fit the single intersection to the dynamically changing traffic flow, and in order to make the single intersection control scheme more accurate. To enhance the accuracy and rationality of the single intersection control scheme, a predictive adaptive control strategy based on K-nearest neighbors for short-term traffic flow is employed. Subsequently, the traffic forecast for adjacent intersections is utilized to determine the appropriate signal cycle. Real-time traffic data is collected to automatically adjust the phase sequence, maximum phase time, and green light extension time. Through thorough simulation analysis using VISSIM software, the Qingdao Huangdao district is selected for single intersection validation. The average vehicle delay time and the ratio of selected green light time for each phase are adopted as evaluation indices. This analysis aims to assess the efficiency of the prediction method for single intersection traffic as well as the prediction method for adjacent intersection traffic.