Purpose: The current study examined the relationship between COVID-mitigation behaviors and mental health through two complementary goals. First, the current study sought to develop classes of individuals, using Latent Class Analysis, based on the different levels of adherence to COVID-mitigation behaviors, and use those classes to predict mental health. Second, the current study sought to develop profiles of individuals, using Latent Profile Analysis, based on various mental health measures, and use those profiles to predict adherence to various COVID-mitigation health behaviors. Participants: Participants included 402 undergraduate students from San Diego State University (M age= 20; 77% women, 21% men). Methodology: Participants were recruited through the SONA participant pool and received course credit or extra credit to complete an online survey. Questions asked about the presence and frequency of COVID-mitigation behaviors (e.g., mask wearing, social distancing) across a variety of essential and non-essential activities. Participants also completed measures of mental health: Satisfaction with Life Scale, Depression, Anxiety and Stress Scale, and Perceived Loneliness Scale. Results: The Latent Class Analyses suggested a 4-class solution: class 1 wore masks but did not distance across all activities, class 2 did not wear masks or distance during non-essential activities, class 3 did not wear masks or distance across all activities, and class 4 wore masks and distanced across all activities. When using the classes to predict mental health, class 2 reported greater anxiety than class 1. The Latent Profile Analyses suggested a 3-profile solution: profile 1 had moderately adjusted mental health, profile 2 had well adjusted mental health, and profile 3 is had maladjusted mental health. When using the profiles to predict adherence, there were no profile differences in COVID-mitigation behaviors. Discussion: The ability to divide individuals into classes based COVID-mitigation behaviors suggests that adherence is nuanced and that intervention efforts may benefit from tailoring to these specific classes.