This review discusses the ferroelectric properties and domain dynamics of AlScN films, their performance optimization through various deposition techniques, and their potential applications in ferroelectric memories and in-memory computing. Ferroelectric materials, which can retain polarization states without external electric fields, are crucial for nonvolatile memory due to their programmable polarization states. However, challenges remain in achieving compatibility with complementary metal-oxide-semiconductor (CMOS) technology and maintaining uniform performance at smaller scales. AlScN, a wurtzite-structured nitride, offers promising solutions due to its high remanent polarization (Pr), low coercive electric field (Ec), and compatibility with CMOS processes. The review highlights the unique advantages of AlScN, including its high Pr and Ec values, stability at high temperatures, and low permittivity, which are beneficial for reducing voltage sharing in non-ferroelectric layers and improving the sensing margin of ferroelectric RAM (FeRAM). AlScN films can be grown using various techniques, including magnetron sputtering, molecular beam epitaxy, and pulsed laser deposition, each with its own advantages and challenges. The performance of AlScN-based ferroelectric memory (FeM) is influenced by factors such as Sc content, in-plane stress, and film thickness. The ferroelectric properties of AlScN are closely related to the existence of metastable layered hexagonal phases, which help reduce the energy barrier between polarization states. The polarization switching in AlScN involves the relative displacement of cations and anions, and the switching process is accompanied by domain nucleation and growth. AlScN-based FeM, including ferroelectric diodes (FeDs), ferroelectric tunnel junctions (FTJs), and ferroelectric field-effect transistors (FeFETs), shows great potential for applications in low-power memory and neuromorphic computing. The review also discusses the challenges and future perspectives of AlScN-based FeM, emphasizing the need for further research to optimize performance and enable commercial applications.This review discusses the ferroelectric properties and domain dynamics of AlScN films, their performance optimization through various deposition techniques, and their potential applications in ferroelectric memories and in-memory computing. Ferroelectric materials, which can retain polarization states without external electric fields, are crucial for nonvolatile memory due to their programmable polarization states. However, challenges remain in achieving compatibility with complementary metal-oxide-semiconductor (CMOS) technology and maintaining uniform performance at smaller scales. AlScN, a wurtzite-structured nitride, offers promising solutions due to its high remanent polarization (Pr), low coercive electric field (Ec), and compatibility with CMOS processes. The review highlights the unique advantages of AlScN, including its high Pr and Ec values, stability at high temperatures, and low permittivity, which are beneficial for reducing voltage sharing in non-ferroelectric layers and improving the sensing margin of ferroelectric RAM (FeRAM). AlScN films can be grown using various techniques, including magnetron sputtering, molecular beam epitaxy, and pulsed laser deposition, each with its own advantages and challenges. The performance of AlScN-based ferroelectric memory (FeM) is influenced by factors such as Sc content, in-plane stress, and film thickness. The ferroelectric properties of AlScN are closely related to the existence of metastable layered hexagonal phases, which help reduce the energy barrier between polarization states. The polarization switching in AlScN involves the relative displacement of cations and anions, and the switching process is accompanied by domain nucleation and growth. AlScN-based FeM, including ferroelectric diodes (FeDs), ferroelectric tunnel junctions (FTJs), and ferroelectric field-effect transistors (FeFETs), shows great potential for applications in low-power memory and neuromorphic computing. The review also discusses the challenges and future perspectives of AlScN-based FeM, emphasizing the need for further research to optimize performance and enable commercial applications.