Stuttering is a speech disorder characterized by disruptions in the fluency of speech, such as repetitions, prolongations, and blocks. Advances in technology, including machine learning and neuro-imaging, are enhancing diagnostic precision and understanding of the disorder’s underlying mechanisms, paving the way for more effective and personalized treatments. This paper provides a thorough review of recent advancements in the field of intelligent processing of stuttered speech. Through an extensive survey of the literature, we explore various approaches ranging from automatic correction and detection to leveraging clinician annotations for improving automatic speech recognition systems. Stuttering diagnosis and classification can be enhanced using machine learning techniques like k-Nearest Neighbors (KNN) and Decision Trees. k-NN classifies speech samples by comparing features such as disfluency frequency and speech rate to labeled instances, identifying patterns indicative of stuttering. Decision Trees, on the other hand, use features like syllable repetitions and silent pauses to create decision rules, providing clear, interpretable classification criteria. These methods improve diagnostic accuracy and enable personalized treatment strategies for stuttering. Confusion metric is used to capture the model’s performance and to showcase achieved results with 89.53% of accuracy with decision tree for word repetition and 86.11% for sound repetition. Stuttering diagnosis and classification face several limitations, including variability in speech patterns, which complicates consistent assessment. Diagnostic tools can be subjective and reliant on clinician expertise, potentially leading to inconsistencies.