Objective: : To investigate the epidemiology of newly-diagnosed, histologically-confirmed (NDHC) central nervous system (CNS) tumours and its changes over a 21-year period in a regional hospital in Hong Kong. Methods: : This is a single-institute retrospective descriptive study of patients undergoing surgery for CNS tumours in a regional hospital of Hong Kong in the period from January 1996 to December 2016. The histological definition of CNS tumours was according to the World Health Organization classification, while the site definition for case ascertainment of CNS tumours was as set out by the Central Brain Tumour Registry of the United States. Patients of any age, who had NDHC CNS tumours, either primary or secondary, were included. The following parameters of the patients were retrieved : age at diagnosis, gender, tumour location, and histological diagnosis. Population data were obtained from sources provided by the Government of Hong Kong. The incident rate, estimated by the annual number of cases per 100000 population, for each histology grouping was calculated. Statistical analyses, both including and excluding brain metastases, were performed. Statistical analysis was performed with Microsoft Excel, 2016 (Microsoft, Redmond, WA, USA). Results: : Among the 2134 cases of NDHC CNS tumours, there were 1936 cases of intracranial tumours and 198 cases of spinal tumours. The annual number of cases per 100000 population of combined primary intracranial and spinal CNS tumours was 3.6 in 1996, and 11.1 in 2016. Comparing the 5-year average annual number of cases per 100000 population of primary CNS tumours from the period 1996–2000 to 2011–2015, there was an 88% increase, which represent an increase in the absolute number of cases by 4.52 cases/100000 population. This increase was mainly contributed by benign histologies. In the aforementioned periods, meningiomas increased by 1.45 cases/100000 population; schwannomas by 1.05 cases/100000 population, and pituitary adenomas by 0.91 cases/100000 population. While gliomas had a fluctuating 5-year average annual number of cases per 100000 population, it only had an absolute increase of 0.51 cases/100000 population between the 2 periods, which was mainly accounted for by the change in glioblastomas. Conclusion : This retrospective study of CNS tumour epidemiology revealed increasing trends in the incidences of several common CNS tumour histologies in Hong Kong, which agrees with the findings in large-scale studies in Korea and the United States. It is important for different geographic locations to establish their own CNS tumour registry with well-defined and structured data collection and analysis system to meet the international standards.