Voice disorders are among the most prevalent conditions today, affecting people globally due to a range of causes including physical, neurological, or functional abnormalities. These disorders can be either temporary or permanent, with symptoms that often define the specific condition, typically diagnosed by a physician through comprehensive evaluation. In recent years, there has been a notable surge in the automation of healthcare services, including disease diagnosis and treatment, driven by advancements in artificial intelligence. Our study explores the application of convolutional neural networks for detecting and predicting voice disorders. To achieve this, we utilized the SVD dataset, comprising 1,516 healthy voices and 1,584 pathological voices, encompassing various phonation types such as low, high, and neutral of /a/, /i/ /u/, as well as sentence. This dataset was used to train and test our proposed methodology. Our objective was to generalize predictions across different voice settings, and we achieved an accuracy of 80.65%, demonstrating the effectiveness of CNNs in diagnosing voice disorders.