NDDs which includes diseases like Parkinson’s, Huntington’s, and Amyotrophic Lateral Sclerosis (ALS) progressively impair neurological function, leading to debilitating symptoms. Gait analysis has emerged as a crucial aspect in the early detection of these diseases, as changes in walking patterns often serve as early indicators of neurodegenerative conditions. The complexity of gait patterns requires sophisticated analysis, making machine learning, particularly deep learning models, essential for accurately identifying subtle changes that may indicate the onset of these diseases. The article introduces a novel deep learning enabled Recurrent Neural Network (RNN) created to detect the existence of Parkinson’s, Huntington’s, and ALS through gaits signals. The model contains 11 layers, an LSTM layer and two dense layers with activation functions that are designed to reflect temporal dynamics of gait data. It was well-performed by this model, especially in the identification of ALS with a sensitivity: 95.64%, specificity: 94.76%, accuracy rate: 95.10%.