•New media use during the outbreak was associated with negative psychological outcomes.•Traditional media use was not associated with psychological outcomes.•Viewing stressful content was associated with more negative affect and depression.•Viewing heroic acts, speeches from experts, and knowledge of the disease and prevention were associated with more positive affect and less depressive symptoms.•Media engagement was associated with more negative affect, anxiety, and stress. The COVID-19 outbreak in China led to an extraordinary threat to public health and wellbeing. This study examined the psychological impact of media use among people indirectly exposed to the disease during the initial phase of the outbreak. We conducted an internet-based survey on January 28, 2020 (one week after the official declaration of person-to-person transmission of the coronavirus). Media use (media forms, content of media exposure, and media engagement) related to the outbreak and psychological outcomes (positive and negative affect, anxiety, depression, and stress) of 917 Chinese adults was assessed. A series of multivariable regressions were conducted. The results showed that use of new media, rather than traditional media, was significantly associated with more negative affect, depression, anxiety, and stress. Viewing stressful content (i.e., severity of the outbreak, reports from hospital) was associated with more negative affect and depression. Media engagement was also associated with more negative affect, anxiety, and stress. However, viewing heroic acts, speeches from experts, and knowledge of the disease and prevention were associated with more positive affect and less depression. The study suggested new media use and more media engagement was associated with negative psychological outcomes, while certain media content was associated with positive psychological impact. The present study highlights the need for timely public health communication from official sources and suggests that reduced exposure to new media may be beneficial.