Aim of this article is to propose a text preprocessing model for sentiment analysis (SA) over twitter posts with the help of Natural Language processing (NLP) techniques. Discussions and investments on health-related chatter in social media keep on increasing day by day. Capturing the actual intention of the tweeps (twitter users) is challenging. Twitter posts consist of Text. It needs to be cleaned before analyzing and we should reduce the dimensionality problem and execution time. Text preprocessing plays an important role in analyzing health-related tweets. We gained 5.4% more accurate results after performing text preprocessing and overall accuracy of 84.85% after classifying the tweets using LASSO approach. © Springer Nature Singapore Pte Ltd. 2019.