Online users’ behaviour and activities are indicated by the opinions expressed on sites. Extremist protest content detection is criticalfor the analysis of the sentiments of users regarding certaingroups and also to deter association with theseillegal actions. Protest content is identified from Twitter tweets in this work. Classification of the Twitter content is performed through feature extraction and their classification. The objective of the Feature Selection techniqueisfor theselection of a relevant and minimal feature subset from a given data set and maintenance of its original representation. Correlation-based Feature Selection (CFS) will assess the value of an attribute subset by taking into account each feature’s predictive capability together with their redundancy degree. Numerous real-world complex problems are handled by the extensive utilization of Biogeography-Based Optimization (BBO). For binary classi?cation problems,the most efficient techniques are Support Vector Machines (SVMs). Arti?cial Neural Networks (ANNs) are the latest and beneficial modelswhich are employedin machine learning and problem-solving. For detecting extremist protestcontent on Twitter, this work examines the proposed FeatureSelection based on BBO.