Huntington's disease (HD) is a neurological condition caused by a trinucleotide repli-cation extension in the Huntington gene. The degeneration of the corticobasal gan-glia's white matter networks causes progressive impairment of motor, cognitive, and neuropsychiatric functioning. Magnetic resonance imaging (MRI) is increasingly be-ing utilized to measure changes in the brain during the early stages of HD, as gene carriers for the disease demonstrate significant neuronal loss until the conclusion of the illness. The absence of tagged data is a significant challenge, particularly in the early phases of the illness. To overcome this, the study's proposed strategy is applied to MRI imaging of Huntington Disease. MRI pictures are pre-processed to reduce noise and improve image quality using normalization. ResNext is a deep learning system that uses MRI pictures to accurately restore hierarchical structures and ex-tract features. After extracting the features, the SNN network is utilized to classify the detection. Finally, by classifying MRI scans as normal, far, mild, near, or HD, the model improves the accuracy of Huntington Disease detection. The proposed HD-RSNN detection accuracy of this optimized system measured at 98.83%. The HD-RSNN datasets were used to validate the suggested approach, yielding accuracy val-ues of 1.81%, 2.40%, and 0.35%, respectively.