The objective of this study is to evaluate characteristics and mortality related to long-term post-stroke pain (PSP). All surviving stroke patients admitted to the Stroke Unit, Haukeland University Hospital, between February 2006 and July 2009 received a postal questionnaire including the fatigue severity scale (FSS), the hospital anxiety and depression scale (HADSD), the Barthel index (BI), and questions regarding location of pain and pain severity at least 6 months after onset of stroke. Survival among patients returning the questionnaire was determined by November 2009. Stroke severity was defined by the modified Rankin score (mRS), 7 days after stroke onset. About 30% of the 408 patients had moderate to severe PSP. On logistic regression, PSP was associated with females (odds ratio (OR) = 2.1, p = 0.002), lower age (OR = 0.98, p = 0.04), fatigue (OR = 3.1, p < 0.001), sleep disturbances (OR = 3.3, p < 0.001), and mRS 3-5 (OR = 1.9, p = 0.03). Among patients with pareses (persistent or transient), there was no difference between paretic and non-paretic side as to frequency of limb pain on follow-up (p = 0.91). By November 2009, 26 patients had died. Cox regression analysis showed that mortality was associated with PSP (hazard ratio (HR) = 2.4, p = 0.040), high age (HR = 1.07, p = 0.001), males (HR = 2.5, p = 0.04), and low BI (HR = 0.97, p < 0.001). In conclusion, our study indicates a multifactorial basis for post-stroke pain. The main new findings were that the frequencies of pain were similar in paretic and non-paretic limbs and that long-term mortality was associated with post-stroke pain.