Abstract Background Hepatocellular carcinoma (HCC) is the most frequent primary invasive cancer of the liver. During the last decade, the epidemiology of HCC has been continuously changing in developed countries, due to more effective primary prevention and to successful treatment of virus-related liver diseases. The study aims to examine survival by level of access to care in patients with HCC, for all patients combined and by age. Methods We included 2018 adult patients (15–99 years) diagnosed with a primary liver tumour, registered in the Palermo Province Cancer Registry during 2006–2015, and followed-up to 30 October 2019. We obtained a proxy measure of access to care by linking each record to the Hospital Discharge Records and the Ambulatory Discharge Records. We estimated net survival up to 5 years after diagnosis by access to care (“easy access to care” versus “poor access to care”), using the Pohar-Perme estimator. Estimates were age-standardised using International Cancer Survival Standard (ICSS) weights. We also examined survival by access to care and age (15–64, 65–74 and ≥ 75 years). Results Among the 2018 patients, 62.4% were morphologically verified and 37.6% clinically diagnosed. Morphologically verified tumours were more frequent in patients aged 65–74 years (41.6%), while tumours diagnosed clinically were more frequent in patients aged 75 years or over (50.2%). During 2006–2015, age-standardised net survival was higher among HCC patients with “easy access to care” than in those with “poor access to care” (68% vs. 48% at 1 year, 29% vs. 11% at 5 years; p < 0.0001). Net survival up to 5 years was higher for patients with “easy access to care” in each age group (p < 0.0001). Moreover, survival increased slightly for patients with easier access to care, while it remained relatively stable for patients with poor access to care. Conclusions During 2006–2015, 5-year survival was higher for HCC patients with easier access to care, probably reflecting progressive improvement in the effectiveness of health care services offered to these patients. Our linkage algorithm could provide valuable evidence to support healthcare decision-making in the context of the evolving epidemiology of hepatocellular carcinoma.