Glycoproteins comprise more than half of current FDA-approved protein cancer markers but the development of new glycoproteins as disease biomarkers has been stagnant. Here we present a pipeline to develop glycoproteins from extracellular vesicles (EVs) through integrating quantitative glycoproteomics with a novel reverse phase glycoprotein array, and then apply it to identify novel biomarkers of breast cancer. EV glycoproteomics show promise in circumventing the problems plaguing current serum glycoproteomics and allowed us to identify hundreds of glycoproteins that have not been identified in serum. We identified 1,453 unique glycopeptides representing 556 glycoproteins in EVs, among which 20 are significantly higher in breast cancer patients. We further applied a novel glyco-specific reverse phase protein array to quantify a subset of the candidates. Together, this study demonstrates the great potential of this integrated pipeline for biomarker discovery.