User handedness can be classified into three primary categories: dominantly left-handed users, dominantly right-handed users, and ambidextrous users. Traditionally, handedness determination is based on observing daily activities such as eating, playing, and writing. However, certain tasks deviate from the usual hand dominance, revealing instances where users’ preferred hand varies or where no dominant hand is evident. This research aims to analyze hand usage in smartphone interactions and identify potential dominance patterns. Users interact with smartphones through various holding patterns, including exclusive use of the left hand, the right hand, or both hands simultaneously. This study classifies these patterns using sensor data collected from accelerometers, gyroscopes, and orientation sensors. Features derived from the sensor data are used with the XGBooster algorithm to classify smartphone handedness into three categories—left hand, right hand, and both hands—achieving an accuracy of 85.04%.