Polyethylene terephthalate (PET), as the largest synthetic polymer with an annual output of more than 90 million tons, has found widespread applications in our daily life and industry, especially its fiber accounts for about 90% of all synthetic fibers. However, PET is highly flammable and melt-dripping during burning, which causes serious casualties and property loss in case of fire. Although many phosphorus-based flame-retardant PETs are already commercially available, their flame retardancy is usually achieved by promoting the melt dripping of PET. The issue on flame retardation of PET without melt dripping was a big challenge both in the academia and industry, and has not been addressed for a long time because the anti-dripping behavior conflicts with the “dripping-promotion” flame-retardant mechanism of phosphorus-containing PET. Therefore, how to flame-retard PET without melt dripping has become an urgent difficult problem needed to solve. The key to designing flame-retardant PET with anti-dripping performance is to increase both the melt viscosity and char formation ability during burning. Herein, we reviewed our original works about a series of flame-retardant and anti-dripping PET copolyesters designed and synthesized via chemically incorporated functional monomers into the PET chains based on “physical interaction”, “high-temperature chemical self-crosslinking” and “high-temperature rearrangement”. Those monomers exhibit high polymerization activity with PET monomers, and those special functional groups do not affect the polymerization or processing process. Once at a higher temperature or during burning, under the actions of “ionic aggregation”, “hydrogen bond”, “chemical self-crosslinking” and “high-temperature rearrangement”, the melt viscosity and melt strength of those PET-based copolyesters would be greatly increased, upgrading excellent anti- or non-dripping performance. What is more, the dense cyclic aromatic structures, enabled by these functional groups, can further evolve into stable intumescent char layers on the surface of copolyesters, and inhibit the heat or oxygen transfer. Thus, the copolyesters exhibit good flame retardancy at the same time. In detail, the physical crosslinking networks, formed by thermally-reversible ionic aggregation, can limit the movement of molecular chains to some extent, but such effects are not enough to make corresponding copolyesters achieve non-dripping performance during the vertical burning test (UL-94). Compared with physical crosslinking, irreversible chemical crosslinking networks are more stable and stronger. Based on this idea, we have proposed a series of novel “intelligent chemical self-crosslinking” copolyesters containing crosslinkable phenylacetylene, phenylacetylene-phenylimide, phenylmaleimide, and aromatic schiff base or azobenzene groups. Those copolyesters can rapidly crosslink during burning or at the temperature higher than polymerization/processing to form stable aromatic cross-linked networks, resulting in excellent flame retardancy and non-dripping properties. In addition, the special aryl ether, bisphenol F and o-hydroxy phenylimide groups can also be used to improve the flame retardancy and anti-dripping of copolyesters via the rearrangement reactions and charring promotion at high temperatures. Among all the above, the self-crosslinkable or rearrangeable PET copolyesters exhibited higher flame-retardant efficiency, which can attain the non-dripping standard at a lower amount of third monomers. Without the use of traditional flame retardant elements such as halogen, these novel mechanisms and approaches have not only opened up a way of high temperature carbonization for the flame-retardant and anti-dripping polyester, but also provided new strategies for the green development of flame-retardant technology. Although the understanding of the design and synthesis of flame-retardant and anti-dripping PET-based copolyesters has been deepened with some breakthroughs in recent years, we still have faced the unavoidable challenges of reducing monomer contents, decreasing cost, optimizing mechanical properties, improving crystallinity and keeping spinnability in the process of practical application and industrialization of these copolyesters in the future. Actually, we have designed some more efficient and cheaper self-crosslinkable monomers recently. The much lower content of monomers can endow corresponding copolyesters with excellent comprehensive performances such as flame retardance, non-dripping, crystallization, spinnability, dyeability and mechanical properties. These state-of-art results will be presented in the near future.