The increasing uses of Internet devices generate significant amount of data. The user communicates with one another using message as there are many people who have the malicious intention to steal our data for the personal gain. Spam can be defined as unwanted data that is being sent to the user. Spam message intended to trick people and to steal the data. This research paper presents an approach to detect the spam message using BERT (Bidirectional Encoder Representation from transformers) algorithm. This study also explores the application of BERT in reference to spam detection, exploring its capabilities, advantages, and performance.