Recently, smartphones evolve, not only used for communication, but also used for more substantial necessity in daily life. Smartphone store more sensitive information such as banking account, or private information about its owner. Due to the needs for safer smartphone security systems, continuous touch behavior authentication is gain significant attention among researchers. However, recent development shown that touch behavior can be compromised by robotic attack. In this paper, we proposed new set of features based on hand stability to authenticate smartphone users. Hand stability are difficult for robotic system to impersonate due to its micro-movement. Our method using hand stability as features for behavioral authentication achieved EER 9.64 % which is potential for future development.