An early and accurate diagnosis of Alzheimer's disease (AD) is required to initiate symptomatic treatment with currently approved drugs and will be of even greater importance if disease modifying compounds in development display a clinical effect. Protein profiles of human cerebrospinal fluid samples from AD patients (n=95) and population-based healthy controls (n=72) were analyzed by SELDI-TOF-MS in order to discover and characterize novel candidate biomarker combinations that differentiate AD patients from normal aging in this explorative study. Thirty candidate biomarkers (ROC AUC>0.7) were discovered that could differentiate patients with AD from healthy controls. Protein sequence determination and positive identification of 15 biomarkers revealed potential associations between the identified markers and AD pathogenesis. A multi-marker combination of five peaks could distinguish AD from healthy control individuals with high sensitivity (97%) and specificity (98%). The panel of five markers was tested on a blinded independent data set of 30 AD samples and 28 controls giving 100% sensitivity and 97% specificity. This novel panel of biomarkers could potentially be used to improve the accuracy of diagnosis of AD.