Although any outcome of an advanced illness can be predicted, in palliative care settings the word ‘prognosis’ usually means the estimated time to death. Prognosis is an important but challenging set of clinical skills for palliative medicine clinicians to master. It is important because patients and families want to know what to expect, it influences clinical decision-making, and it may determine eligibility for services. It is challenging because of the inherent uncertainty of making predictions and because dying is not an easy topic to discuss. Advances in statistical computing have allowed the development of mathematical models and predictive tools that are now more accurate than clinical estimates. A large section of this chapter is devoted to presenting and evaluating several of these models, although prognostic uncertainty remains a significant issue even with them, and survival estimates should never drive clinical decision-making alone.