Abstract: To explore the latent profiles of physical activity among junior and senior high school students and the association with short-video addiction, so as to provide evidence for the prevention of short-video addiction among adolescents. Methods: In November 2024, a stratified random sampling method was used to select 20 840 students from 120 junior and senior high schools in Guangdong Province. Physical activity and short-video addiction were assessed using International Physical Activity Questionnaire and Short-video Addiction Scale. Latent profile analysis was used to identify latent classes of physical activity. The robust three-step approach was used to examine the effects of demographic variables, including sex, school stage, age, and residence, on profile membership, and the Bolck-Croon-Hagenaars (BCH) method was used to compare short-video addiction levels across profiles after adjusting for demographic variables. Results: Four latent profiles of physical activity in junior and senior high school students were identified:low-frequency low-duration profile ( n =1 668, 8.00%), weekday high-frequency activity profile ( n = 12 408 , 59.54%), weekday low-frequency activity profile ( n =5 107, 24.51%), and high frequency high-duration profile ( n = 1 657 , 7.95%). With the high-frequency high-duration profile as the reference, girls were more likely to be assigned to the weekday high-frequency activity profile and less likely to be in the low-frequency low-duration profile compared with boys ( B =0.27, -0.33 ); senior high school students had higher odds of membership in the other three profiles than junior high school students ( B =0.63- 1.09 ); increasing age was associated with increased odds of the low-frequency low-duration profile ( B =0.11); rural students were less likely to belong to the other three profiles than urban students ( B =-0.97 to -0.79) (all P <0.01). After adjusting for demographic variables, short-video addiction scores differed significantly across physical activity profiles in junior and senior high school students ( χ 2=390.23, P <0.01), with scores decreasing in the following order:low-frequency low-duration profile, weekday low-frequency activity profile, weekday high-frequency activity profile, and high-frequency high-duration profile. Conclusions Short-video addiction levels vary across different physical activity patterns in junior and senior high school students. Overall, participants with high-frequency and high-duration participation score the lowest, whereas those with low-frequency and low-duration participation score the highest.