In less than a decade, metal halide perovskites (MHPs) have been demonstrated as promising solar cell materials because the photoelectric conversion efficiency (PCE) of the representative material CH3NH3PbI3 rapidly increased from 3.8% in 2009 to 25.2% in 2009. However, defects play crucial roles in the rapid development of perovskite solar cells (PSCs) because they can influence the photovoltaic parameters of PSCs, such as the open circuit voltage, short-circuit current density, fill factor, and PCE. Among a series of superior optoelectronic properties, defect tolerance, i.e., the dominate defects are shallow and do not act as strong nonradiative recombination centers, is considered to be a unique property of MHPs, which is responsible for its surprisingly high PCE. Currently, the growth of PCE has gradually slowed, which is due to low concentrations of deep detrimental defects that can influence the performances of PSCs. To further improve the PCE and stability of PSCs, it is necessary to eliminate the impact of these minor detrimental defects in perovskites, including point defects, grain boundaries (GBs), surfaces, and interfaces, because nonradiative recombination centers seriously affect device performance, such as carrier generation and transport. Owing to its defect tolerance, most intrinsic point defects, such as VI and VMA, form shallow level traps in CH3NH3PbI3. The structural and electronic characteristics of the charged point defect VI − are similar to those of the unknown donor center in a tetrahedral semiconductor. It is a harmful defect caused by a large atomic displacement and can be passivated to strengthen chemical bonds and prevent atom migration by the addition of Br atoms. Owing to the ionic nature of MHPs and high ion migration speed, there are a large number of deep detrimental defects that can migrate to the interfaces under an electric field and influence the performance of PSCs. In addition, the ionic nature of MHPs results in surface/interface dangling bonds terminated with cations or anions; thus, deep defects can be passivated through Coulomb interactions between charged ions and passivators. Hence, the de-active deep-level traps resulting from charged defects can be passivated via coordinate bonding or ionic bonding. Usually, surface-terminated anions or cations can be passivated by corresponding cations or anions through ionic bonding, and Lewis acids or bases can be passivated through coordinated bonding. In this review, we not only briefly summarize recent research progress in defect tolerance, including the soft phonon mode and polaron effect, but also strategies for defect passivation, including ionic bonding with cations or anions and coordinated bonding with Lewis acids or bases.