In today's technologically sound world the use of social media is inevitable. Along with benefits of social media there are serious negative impacts as well. An important issue that needs to be addressed here is cyberbullying. An effective solution for resolving this issue is the detection of the cyber-bullying content by Machine Learning. This manuscript aims to put forward ideas regarding cyber-bullying detection on the social media platform twitter. The outcome of this manuscript is that whichever tweet is a bully tweet that is represented by the value 1, thus all the bully tweets are detected. The Twitter dataset is equally distributed into bully and non-bully tweets and fed to different machine learning models. The logistic regression classifier provides accurate classification of bully and non-bully tweets with precision of 91%, recall 94% and F1-score 93%. This work will help curb cyber-bullying, so that the users can stay at bay from victimization.