Cancer stem cells (CSCs) were isolated in multiple tumor types, including human glioblastomas, and although the presence of surface markers selectively expressed on CSCs can be used to isolate them, no marker/pattern of markers are sufficiently robust to definitively identify stem cells in tumors. Several markers were evaluated for their prognostic value with promising early results, however none of them was proven to be clinically useful in large-scale studies, leading to outstanding efforts to identify new markers. Given the heterogeneity of human glioblastomas further investigations are necessary to identify both cancer stem cell-specific markers and the molecular mechanisms sustaining the tumorigenic potential of these cells to develop tailored treatments. Markers for glioblastoma stem cells such as CD133, CD15, integrin-α6, L1CAM might be informative to identify these cells but cannot be conclusively linked to a stem cell phenotype. Overlap of expression, functional state and morphology of different subpopulations lead to carefully consider the techniques employed so far to isolate these cells. Due to a dearth of methods and markers reliably identifying the candidate cancer stem cells, the isolation/enrichment of cancer stem cells to be therapeutically targeted remains a major challenge.