The ICD-10 code provides accurate and updated procedural codes for the improvement of health care diagnosis, cost and ensures an im-partial reimbursement policies. ICD-10-CM is followed and implemented internationally to provide a quality health care for the patients on a global scale. The clinical environment knowledge in a natural language form detects each sentence. In order to maintain positivity, remove all the negative words in the sentence. Dependent clause that provides a sentence element with additional information and which cannot stand alone in a sentence are identified and removed. The resultant sentence is then preprocessed using Text mining techniques. The extracted meaningful words are then processed through the available huge volume of ICD-10 CM codes database. The main aim of this paper is to map the perception of complaints onto an abstract representation and reasoning the system to generate an appropriate ICD-10 CM code. The idea of the work is to provide efficiency on complex vocabulary, vague and imprecise terms, synonymy and polysemy terms. The effectiveness of this proposed work is determined through the process of Perceptron for finding the efficiency between the trained and test dataset. © BEIESP.