. Objectives.  To compare lipoprotein components associated with ischaemic and haemorrhagic stroke by age and gender in the Apolipoprotein MOrtality RISk (AMORIS) Study (n = 148 600). Design.  Prospective follow‐up study (11.8, range 7–17 years) of fatal and nonfatal ischaemic and haemorrhagic stroke through linkage with Swedish hospital discharge and mortality registers. Setting.  Measurements of lipoprotein components from health check‐ups in the larger Stockholm area. Results.  Ischaemic stroke was more common than haemorrhagic stroke (5 :1), but case fatality was higher in haemorrhagic stroke. An elevated apoB/apoA‐1 ratio and triglycerides, non‐HDL cholesterol, low HDL cholesterol, and the total cholesterol to high‐density cholesterol (TC/HDL‐C) ratio were associated with increased incidence of nonfatal and fatal ischaemic stroke as well as all cerebrovascular events (n = 7480) in both genders. The associations were somewhat stronger for nonfatal than fatal events. In ischaemic stroke the apoB/apoA‐1 ratio was a stronger predictor than the TC/HDL‐C ratio in all subjects, in those below 65 years of age and in those with LDL‐C below 3 mmol L−1. Haemorrhagic stroke was not associated with elevated atherogenic lipoproteins except for increased risk of fatal haemorrhagic stroke in women with a high apoB/apoA‐I ratio. Conclusions.  Dyslipidaemia is associated with an increased risk of ischaemic stroke but few relations were seen in haemorrhagic stroke. Use of the apoB/apoA‐I ratio as a marker of dyslipidaemia is at least as efficient as conventional lipids, for the identification of subjects at increased risk of stroke, especially ischaemic stroke. Practical advantages, fasting is not needed, speak in favour of using apoB and apoA‐1 in stroke risk prediction.