Parkinson’s disease is a neurological disorder that worsens over time which cannot be cured but with the help of medicines, symptoms can get better. Neurological disorders mainly affect the motor and cognitive skills of individuals. Keystroke Dynamics which came into existence in the late 19th century can be a prominent solution to detect the severity of the disease making the treatment more efficient. Keystroke Dynamics is capable of measuring unique features of users like key press duration, typing rate in milliseconds precision which help to detect the severity of the disease. The objective of this paper is to detect the accuracy of severity of Parkinson’s Disease of the affected individuals. Various machine learning algorithms like Naive Bayes’ algorithm, Support Vector Machine, K-Nearest Neighbor, Decision Tree, Random Forest were used to detect the accuracy of severity. In addition, we have also built an android app which can collect keystroke information of the user such as their Hold Time, Press-Press Latency, Release-Release Latency, Flight Time, Correction Duration, Pre-Correction Slowing, PostCorrection Slowing, and After Punctuation Pause. These information will help to detect when they are suffering from Parkinson’s Disease and how severe it is.