Smartphones have become an integral part of students’ daily lives; however, excessive and unconscious usage can impact their productivity and mental well-being. Traditional survey-based studies are often limited by self-reporting inaccuracies, leading to the need for more objective assessment methods. This study adopts a smartphone sensing approach rooted in digital phenotyping to collect real-time behavioral data from 30 university students. Using an in-house developed app, data on app usage, screen unlock patterns, sleep indicators, and stress levels were collected. Behavioral features like mobile usage, sleep quality, and stress levels were used to predict distracted and mindful usage windows. Findings indicate that high non-academic app usage during study hours, frequent nighttime phone activity, and elevated stress levels correlate with distracted behavior. Conversely, healthier sleep patterns and the use of academic-related apps are associated with mindful engagement. These insights aim to promote greater self-awareness and healthier smartphone habits among students.