Modern neuroimaging safely and reliably diagnoses stroke and provides information for outcome prediction. However, some patients with clinical stroke have no detectable abnormality on neuroimaging and other patients are not fit for such investigations. Therefore, we evaluated the potential of plasma DNA and serum S100 protein concentrations to predict post-stroke mortality and morbidity in patients with negative neuroimaging results. Patients with stroke-like symptoms but negative neuroimaging results were recruited. Both plasma and serum were collected from each patient for plasma DNA and serum S100 analysis. The primary outcome measures were 6-month mortality and morbidity using the post-stroke modified Rankin score (mRS). Forty-four patients were recruited to the study. Seventeen (39%) patients were classified as post-stroke mRS grades 3-6. The median plasma DNA concentration of this group of patients was significantly higher than that of patients with post-stroke mRS grades 0-2. Median serum S100 protein concentrations did not show significant differences between the two groups. Plasma DNA concentrations > 800 kilogenome-equivalent/l have a sensitivity of 42% and a specificity of 100% for predicting 6-month post-stroke mRS (grades 0-2), with an area under the receiver operator characteristic (ROC) curve of 0.742. By comparison, serum S100 protein concentrations > 0.09 microg/l have a sensitivity of 48% and specificity of 75% for predicting 6-month post-stroke mRS (grades 0-2), and the area under the curve is 0.542. Plasma DNA concentration predicts post-stroke morbidity and mortality in patients with negative neuroimaging, and may be more effective than S100 protein measurement.