This study focuses on discovering factors associated with smartphone addiction (SA) among high school students. The study was conducted by surveying 14,036 high school students on demographic characteristics as well as the Middle School Student Mental Health Scale (MSSHS). This scale gathered information to identify characteristics such as social anxiety, smartphone addiction, shame-proneness, loneliness, and peer victimization, and possible signs of SA. Then, the data was run through nine machine learning algorithms with the XGBoost algorithm achieving the highest predictive ability. After analyzing the testing set, which included 70% of surveyed students, the study found that 36.1% of students displayed SA. Higher scores in shame-proneness, learning pressure, social anxiety, mood swings, interpersonal sensitivity, peer victimization, anxiety, and the presence of non-suicidal self-injury were predictive of SA. One of the main points discussed was how the results of this study display the value of combining predictive modeling with network analysis to identify SA-related factors, helping with providing a basis for targeted interventions.