In recent years, traffic density forecasting has been playing an important role in developing and improving the performance of intelligent traffic systems. Traffic density forecasting to optimize traffic management, urban management of vehicular traffic have the ability to coordinate traffic, optimize traffic light signals and apply intelligent regulation based on forecasts, thereby improving the handling of the road segment and reducing travel time. Therefore, the construction of predictive algorithms along with integration into traffic management systems is essential to promote the sustainable development of intelligent transportation systems in the future. In this research, an algorithm has been developed to predict traffic density on the delayed Lighthill - Whitham - Richards (LWR) model, in which a regularization difference method has been proposed as the basis for the algorithm. This article mainly focus on building an algorithm and performing experimental calculations to verify the correctness of the algorithm on a mathematical model.