Chatbot is a popular application used to converse with a computer-based system in an interactive and amicable manner. It enables a person to interact with the vast amount of knowledge in an easier and cost-effective way. Chatbots are able to answer the questions using its knowledge base just like a human expert, the only exception is that it takes mere minutes or even seconds for a chatbot to learn as opposed to the human's ability to acquire the same knowledge over a period of months or years. The learning rate depends on the algorithm applied, processing speed of the system and the amount of data to train the model. This particular characteristic of quick learning proves valuable in all application domains and more so in the health services. In this paper, we propose a health chatbot developed using deep learning techniques of complex neural networks. The model is trained on a Knowledge Base which consists of health queries and its answers. These question-answer pairs were obtained manually from WebMD, which is one of the leading health service websites on the internet. An accuracy of 0.97% and loss of 0.02%, in terms of correct answer prediction for a given query, was obtained while training the deep learning model.