<p>Accurate short-term traffic flow forecasting is crucial for efficient traffic management and intelligent transportation systems. Understanding future traffic trends through short-term predictions can significantly enhance proactive traffic management strategies. This paper proposes a novel deep learning-based prediction model that leverages Variational Mode Decomposition (VMD) in conjunction with Convolutional Neural Network (CNN) and Long Short-Term Memory Network (LSTM) to enhance the accuracy and reliability of traffic flow forecasting. The model, referred to as VMD-CNN-LSTM, begins with VMD to decompose complex traffic flow data into several intrinsic mode functions, isolating the intrinsic patterns and reducing noise. Each intrinsic mode function is then separately modeled using a CNN to capture hierarchical temporal features and nonlinear relationships. Finally, the outputs of the CNNs are fed into an LSTM network, which processes the sequential data in both forward and backward directions, capturing long-term dependencies and context. Extensive experiments conducted on real-world traffic datasets demonstrate that our VMD-CNN-LSTM model significantly outperforms traditional methods in terms of prediction accuracy and robustness.</p>