Abstract: Despite the marked success of applications of PD-1/PD-L1 checkpoint blockades in clinical practice, the efficacy and responsiveness of these agents varies greatly among different tumor types and across individual patients. Therefore, establishment of predictive biomarkers for checkpoint blockades is of great importance to maximize the therapeutic benefits. In this review, we discuss the current progress and challenges of developing predictive biomarkers of immunotherapy responsiveness, aiming to provide directions for future studies. PD-L1 expression is a logical biomarker for the prediction of response to anti-PD-(L)1 immunotherapies.However, the predictive values of PD-L1 expressions for immunotherapy still remain controversial. Multiplex detecting methods and combined biomarkers may provide new strategies, including tumor mutation burden, some oncogene mutations, like EGFR, ALK, KRAS and STK11. As current evidence of those potential predictors, a consensus and standardization is needed to be established for popularization in future studies.