The COVID-19 pandemic has led to a rise in opposition to vaccination, hindering herd immunity. As social media play a major role in the formation of anti-vaccination communities, it is critical to monitor the discourse on the platforms to effectively counter the negative sentiment and encourage people to vaccinate. This study employs computational content analysis, specifically topic modeling and time series analysis, to monitor the COVID-19 anti-vaccination communities on Facebook in Czechia. The analysis generated 18 topics with politics, governance, and international affairs being the most discussed, and only five dealt with issues directly related to COVID-19. Discussions about information and its credibility were prevalent, and members of these anti-vaccination communities relied heavily on social media content and conspiracy websites as sources of information, while neglecting scientific resources. The study highlights the need for ongoing monitoring of anti-vaccination communities on social media and the development of effective communication strategies to promote vaccination.