Silicon heterojunction (SHJ) solar cells are crystalline silicon wafer-based photovoltaic devices fabricated with thin-film deposition technology. The SHJ solar cells hold great potential for large-scale deployment for high conversion efficiencies with low-cost manufacturing. Recently Kaneka Corporation has fabricated an interdigitated-back-contact (IBC) SHJ solar cell with a certified 26.33% conversion efficiency in a large area (180.4 cm2), which is a world record for any 1-sun crystalline silicon wafer-based solar cell. The key feature of SHJ solar cells is the impressive highopen-circuit voltages (Voc) achieved by the excellent amorphous/crystalline silicon interface passivation. Generally, in SHJ solar cells, the boron doped hydrogenated amorphous silicon [(p)a-Si:H] serves as hole collector and the phosphorus doped hydrogenated amorphous silicon [(n) a-Si:H] functions as electron collector. In order to improve the lateral carrier transport of these layers, transparent conductive oxides (TCOs) are usually deposited on both sides of the solar cell. Therefore the parameters such as the heterointerface passivation quality, doping concentration and thickness of the a-Si:H doped layer, and work function of the transparent conductive oxide layer are the key factors that affect the performances of SHJ solar cells. Enormous research efforts have been devoted to studying the effects of the aforementioned influencing parameters on the photovoltaic characteristics of SHJ solar cells. Some research groups have addressed the physical mechanism behind the limitation of the solar cell efficiency. Owing to the insight into the physical mechanism some guidelines for optimally designing the high-performance solar cells in future are obtained. It seems therefore important to summarize the research efforts devoted to the physical mechanism and optimal design of SHJ solar cells.In the present review, we mainly discuss three important issues: 1) the amorphous/crystalline silicon interface passivation; 2) the Schottky barrier resulting from the work function mismatch between the (p)a-Si:H doped layer and the transparent conductive oxide layer; 3) the screening length that is required to efficiently shield the parasitic opposing band from bending originating from the work function mismatch between the (p)a-Si:H doped layer and the transparent conductive oxide layer. The numerical simulation and optimal design of SHJ solar cells are analyzed, and three strategies that may improve the solar cell performances are presented: 1) a hybrid SHJ solar cell structure with a rear heterojunction emitter and a phosphorus-diffused homojunction front surface field; 2) replacing the (p)a-Si:H doped layer by higher doping efficiency microcrystalline silicon alloys such as c-Si:H, c-SiOx:H or c-SiCx:H; 3) replacing the (p)a-Si:H doped layer by higher work function transition metal oxides such as MoOx, WOx or VOx. Finally, the research progress and future development of SHJ solar cells are also described.