The 5-year relative survival rate of ovarian cancer is higher than 90% if detected early, whereas drastically lowered to under 20% when detected after stage III. Early diagnosis is necessary to improve treatment outcome and survival rate of the ovarian cancer. As ovarian cancer biomarkers, serum CA125 could be detected alone. It has been known that serum and salivary CA125 were correlated in ovarian cancer patients. Therefore, we developed phage-based biosensor to detect the CA125 ovarian cancer biomarker in saliva. We identified highly specific phages to CA125 through phage display screening. We then genetically engineered the phage to express CA125 receptors on the major coat proteins of phages. Using the engineered phage, we fabricated self-assembled matrices composed of quasi-ordered fiber bundle structures to exhibit tunable colors by self-templating assembly approach. Upon exposure of different concentrations of biomarker, the multi-color phage matrices exhibited distinct color changes that can be correlated to colorimetric measurement. Furthermore, we used home-made smartphone app to detect the biomarker selectively and sensitively from the human saliva. Our phage-based ovarian cancer biomarker detection approach will be useful for early detection of ovarian cancer and increase of the survival rate because it is simple, convenience, and making self-diagnosis possible.