Millions of individuals are living with voice disorders which influence their communication skills, socialization and overall wellbeitableng. Early diagnosis and accurate diagnosis is crucial in providing relevant interventions and treatment of various illnesses. Over the last few years, machine learning has received promising results in several healthcare applications, including the diagnosis of vocal disorders. This paper is a comprehensive assessment and comparison of machine learning in the detection of voice disorders. The paper start with the various categories and etiologies of voice disorders with a focus on the need to have proper and effective means of detecting the disorder and discussed different machine learning algorithms used in this field, such as neural networks,randaom forest,support vector machine and deep learning architecture. We discuss the challenges associated with feature selection and extraction, as well as the limitations of existing datasets, highlighting the necessity of more extensive and varied datasets in order to train reliable models. Additionally, we compare the performance of different machine learning algorithms and evaluate their accuracy, sensitivity, specificity, and overall diagnostic capabilities. We analyse the strengths and weaknesses of each approach, identifying potential areas for improvement and future research directions.