Bot detection is a term that refers to the ability of behavioral biometrics to differentiate between humans and bots. Human interactive proofs (HIPs), such as CAPTCHA, are the current approaches for distinguishing between humans and bots. Because most CAPTCHA systems can be solved using state-of-the-art AI approaches, CAPTCHAs are not perfect for bot detection. Bots are getting increasingly intelligent, with the ability to imitate human behavior. In this paper, the proposed method to distinguish between human and bot is based on endpoint prediction of mouse movement utilizing behavioral biometrics and mouse dynamics. First collected human and bot mouse movement data, after data collection, split this data into training and testing parts and build the template libraries, matching these templates libraries with each other and distinguishing between human and bot on the base of endpoint prediction. Different types of experiments were performed to distinguish between human and bot. First, use human data to train prediction model and evaluate prediction performance on both human and bot testing templates and the target hit rate is 10.6% on human testing (Human-Human) templates and 3.5% on bot testing (Human-Bot) templates when 90% of movement is completed. After that use bot data to train prediction model and evaluate prediction performance on both bot and human testing templates and the target hit rate is 11.2% on bot testing (Bot-Bot) templates and 5.2% on human testing (Bot-Human) templates when 90% of movement is completed. Overall, the target hit rate can easily discriminate between human and bot.