Gait abnormality classification of neurodegenerative disorders plays a vital role in early diagnosis and clinical decision-making. Traditional machine learning approaches often rely on handcrafted features and fail to capture the complex temporal and spatial dynamics present in gait signals, limiting their classification performance. To overcome these, we present a hybrid deep learning model combining Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Gated Recurrent Units (GRU) for classifying gait patterns associated with neurodegenerative diseases using the PhysioNet GaitNDD dataset. The model achieves high classification accuracies of 97.14% for NDD vs HC, 98.06% for ALS vs HC, 98.14% for HD vs HC, and 96.22% for PD vs HC, outperforming existing methods. A detailed explainability analysis using SHAP highlights the importance of features such as sample entropy and stride interval, confirming the model’s focus on clinically relevant patterns.