Abstract Background Early palliative care (EPC) in oncology has been shown to have a positive impact on clinical outcome, quality-of-care outcomes, and costs. However, the optimal way for activating EPC has yet to be defined. Methods This prospective, multicentre, randomised study was conducted on 207 outpatients with metastatic or locally advanced inoperable pancreatic cancer. Patients were randomised to receive ‘standard cancer care plus on-demand EPC’ (n = 100) or ‘standard cancer care plus systematic EPC’ (n = 107). Primary outcome was change in quality of life (QoL) evaluated through the Functional Assessment of Cancer Therapy – Hepatobiliary questionnaire between baseline (T0) and after 12 weeks (T1), in particular the integration of physical, functional, and Hepatic Cancer Subscale (HCS) combined in the Trial Outcome Index (TOI). Patient mood, survival, relatives' satisfaction with care, and indicators of aggressiveness of care were also evaluated. Findings The mean changes in TOI score and HCS score between T0 and T1 were −4.47 and −0.63, with a difference between groups of 3.83 (95% confidence interval CI 0.10–7.57) (p = 0.041), and −2.23 and 0.28 (difference between groups of 2.51, 95% CI 0.40–4.61, p = 0.013), in favour of interventional group. QoL scores at T1 of TOI scale and HCS were 84.4 versus 78.1 (p = 0.022) and 52.0 versus 48.2 (p = 0.008), respectively, for interventional and standard arm. Until February 2016, 143 (76.9%) of the 186 evaluable patients had died. There was no difference in overall survival between treatment arms. Interpretations Systematic EPC in advanced pancreatic cancer patients significantly improved QoL with respect to on-demand EPC.