Purpose: Travel websites allow tourists to share their thoughts, beliefs and experiences regarding various travel destinations. In this paper, the researchers demonstrated an approach for destination marketing organisations to explore online tourist-generated content and understand tourists' perceptions of the destination image (DI). Specifically, the researchers initiated an investigation examining how the destination image of Macau changed during the period of 2014–2018 based on user-generated content on travel websites. Design/methodology/approach: Web crawlers developed by Python were employed to collect tourists' reviews from both Ctrip and TripAdvisor regarding the theme of “Macau attraction”. A total of 51,191 reviews (41,352 from Ctrip and 9,839 from TripAdvisor) were collected and analysed using the text-mining technique. Findings: The results reveal that the frequency of casino-related words decreased in reviews by both international and mainland Chinese tourists. Additionally, international and mainland Chinese tourists perceive the DI of Macau differently. Mainland Chinese tourists are more sensitive to new attractions, while international tourists are not. The study also shows that there are differences between the government-projected DI and the tourist-perceived DI. Only the “City of Culture” and “A World Centre of Tourism and Leisure” have built recognition with tourists. Originality/value: Given the easy accessibility of online information from various sources, it is important for destination marketing organisations to analyse and monitor different DI perspectives and adjust their branding strategies for greater effectiveness. This study uncovered the online DI of Macau by using text mining and content analysis of two of the largest travel websites. By analysing and comparing the differences and relationships among the frequently used words of tourist-generated content on these websites, the researchers revealed some interesting findings with important marketing implications.