ConspectusUtilizing light to enable chemical conversions presents a green and sustainable approach to produce fuels and chemicals, and photocatalysis is one of the key chemical technologies that needs to be well developed in this century. Despite continuous progress in the advancement of various photocatalysts based on small inorganic and organic compounds, polymers, and networks, designing and constructing photocatalysts that combine activity, selectivity, and reusability remains a challenging goal. For catalytic activity, the difficulty originates from the complexity of photochemical reactions, where the light-harvesting system, multielectron and multihole-involving processes, and pinpoint mass delivery simultaneously need to be established in the system. For selectivity, the difficulty stems from the elaborate design of catalytic sites and space, especially their orbital energy levels, spatial arrangement, and environment; developing a molecular strategy that enables an overall design and control of these factors of different aspects is necessary yet arduous. For reusability, the difficulty arises from the stability and recyclability of the photocatalysts upon continuous operation under photoredox reaction conditions. How to recover photocatalysts in an energy-saving way to enable their cyclic use while retaining activity and selectivity is at the core of this problem. These bottleneck issues reflect that molecular design of a photocatalyst is not a simple summation of the above requirements, but a systematic scheme that can organically interlock various aspects is needed.To enable such an elaborate design and precise control, a basic requirement of the scaffold for constructing a promising photocatalyst is that its primary and high-order structures should be molecularly predesignable and synthetically controllable. Such a molecular regime has successfully evolved in natural photosynthesis, where light-harvesting chlorophyll antennae and photocatalytic centers are spatially well-organized and energetically well-defined to build ways for exciton migration, photoinduced electron transfer and charge separation, electron and hole flows, and oxidation of water and reduction of carbon dioxide, thereby converting water into oxygen to release ATP and NADPH via the light reaction and carbon dioxide into glucose with ATP and NADPH through the dark reaction. Similarly, a predesignable polymeric scaffold would be promising for integrating these complex photochemical processes to construct photocatalysts.Covalent organic frameworks (COFs) are a class of extended yet polymeric materials that enable the organization of organic units or metallo-organic moieties into well-defined architectures. In principle, COFs are molecularly designable with topology diagrams and synthetically controllable through polymerization reactions, offering an irreplaceable platform for designing and synthesizing photocatalysts. This feature enticed researchers to develop various photocatalysts based on COFs and drove the rapid progress in this field over the past decade. In this Account, we summarize the recent advances in the molecular design and synthetic control of COF photocatalysts, by highlighting the key achievements in developing ways to enable light harvesting, trigger photoinduced electron transfer and charge separation, allow charge carrier transport and mass delivery, control energy level, catalytic space, and environmental engineering, and develop stability and recyclability with an aim to reveal a full picture of this field. By scrutinizing typical photocatalytic reactions, we show the key problems to be addressed for COFs and predict future directions.