Semiconductor-based photocatalysis provides a sustainable and green pathway for pursuing hydrogen peroxide (H2O2 ) production, showing great potential in the fields of pollutant degradation, chemical synthesis, and phototherapy. Compared with other photocatalytic reactions (like carbon dioxide reduction, nitrogen fixation, and water splitting), photocatalytic H2O2 production via molecular oxygen (O-2) reduction and/or water (H2O) oxidation is more feasible in both thermodynamics and kinetics aspects. Moreover, the high chemical reactivity of H2O2 also endows the system with unique performance-evaluation methods and application scenarios. The above characteristics complicate the relevant investigations, and the photocatalyticsystem design deserves more special attention. Recently, benefiting from the development of synthetic and characterization techniques, great progress has been made in semiconductor-based photocatalytic H2O2 production. In this work, we provide an overview of recent advances in the field of semiconductor-based photocatalytic H2O2 production. In the first part, we summarize the mechanism of semiconductor-based photocatalytic H2O2 production. Typically, photocatalytic H2O2 production could undergo a series of reaction pathways, mainly involving electron-transfer-mediated O2 reduction and hole-transfer-mediated H2O oxidation. The versatile reaction pathways, on the one hand, make it difficult to quantitate the contributions of different processes, on the other hand, enable diverse optimization methods to be established. The characterization techniques for interrogating excited-state and structural properties involved in the reaction are also summarized. Moreover, we discuss the general optimization methodologies for pursuing advanced photocatalytic H2O2 production. To go further, we summarize recent progress in the performance evaluation of photocatalytic H2O2 production. Owing to the relatively high reactivity of H2O2 and the requirement of product separation, the reactor has been demonstrated to impact the performance of H2O2 production, where a continuous flow reactor has been widely employed in evaluating the photocatalytic performance. A series of methods for H2O2 detection are summarized, and their advantages and disadvantages are discussed. Besides, the key parameters for assessing the performance of photocatalytic H2O2 production including production rate, apparent quantum yield, and solar-to-chemical conversion efficiency are summarized. Then, we summarize the general materials for triggering photocatalytic H2O2 production, where metal oxides, metal sulfides, polymers, metal-organic frameworks, and covalent organic frameworks are presented. We discuss their advantages and disadvantages in aspects including chemical stability, light absorption, photogenerated charge carrier separation, and so on. The progress in these catalyst systems is summarized, where the general optimization strategies and photocatalytic performance are presented. In the fourth part, we present the application scenarios of semiconductor-based photocatalytic H2O2 production. Compared with other traditional methods for H2O2 synthesis, the photocatalytic one usually gives rise to a relatively low H2O2 yield. Such low-yield H2O2 production not only raises requirements for developing product-enrichment strategies but also inspires researchers to pay attention to those application scenarios with low H2O2 concentration demand. In this context, the applications of semiconductor-based photocatalytic H2O2 production are mainly limited to those scenarios in pollutant treatment, sterilization, and phototherapy. In addition, the coupling between photocatalytic H2O2 production and oxidative organic transformations is also getting considerable attention by virtue of its advantage in gaining concomitant generation of high-value chemicals and boosting the yield of H2O2 generation. At the end of the review, we discuss several major challenges in semiconductor-based photocatalytic H2O2 generation. The challenges in mechanism investigation (like reactive-site identification, reaction-pathway recognition, surface structure reconstruction, etc.), standard methods for photocatalytic performance evaluation (like reactor, solvent, and atmosphere), and application scenario exploration are presented, followed by a brief discussion of the possible solutions.