This study aimed to assess the mental health of a University community in South Brazil during the COVID-19 pandemic. A cross-sectional web-based survey was conducted between July-August 2020 through a self-administered questionnaire. All University staff and students were eligible. Depression was measured using the Patient Health Questionnaire-9 and anxiety by the Generalized Anxiety Disorder-7. To evaluate the effect of social distancing and mental health factors on outcomes, Poisson regression models with robust variance were performed, estimating Prevalence Ratios (PR) and 95% Confidence Intervals (95%CI). 2,785 individuals participated in the study. Prevalence of depression and anxiety were 39.2% (95%CI 37.3-41.1) and 52.5% (95% CI 50.6-54.4), respectively. Undergraduate students showed a higher prevalence of the outcomes. Not leaving the house routinely, mental health care, and previous diagnosis of mental illness were associated with both outcomes. Those with a previous medical diagnosis of depression had a 58% (PR 1.58; 95%CI 1.44; 1.74) and anxiety a 72% (PR 1.72; 95%CI 1.56; 1.91) greater prevalence of depression than their peers. An alarming prevalence of psychopathologies was observed. Despite the well-known benefits of social distancing to public health, it requires a surveillance on the population's mental health, especially students and those with previous mental illness diagnosis.