As cloud computing and massive-data machine learning are applied pervasively, ultra-high volume data storage serves as the foundation block. Every day, nearly 2.5 quintillion bytes (50000 GB/second in 2018) of data is created and stored. Hard Disk Drive (HDD) takes major part of this heavy duty. However, despite the amazing evolution of HDD technology during the past 50 years, the conventional Perpendicular Magnetic Recording (PMR), the state-of-the-art HDD technique, starts to have less momentum in increasing storage density because of the recording trilemma. To overcome this, Heat-Assisted Magnetic Recording (HAMR) was initially proposed in 1990s. With years of advancement, recent industrial demos have shown the potential of HAMR to actually break the theoretical limit of PMR. However, to fully take advantage of HAMR and realize the commercialization, there are still quite a few technical challenges, which motivated this thesis work. Via thermal coupled micromagnetic simulation based upon Landau-Lifshitz-Bloch (LLB) equation, the entire dynamic recording process has been studied systematically. The very fundamental recording physics theorem is established, which manages to elegantly interpret the previously conflicting experimental observations. The thermal induced field dependence of performance, due to incomplete switching and erase-after-write, is proposed for the first time and validated in industrial lab. The combinational effects form the ultimate physical limit of this technology. Meanwhile, this theorem predicts the novel noise origins, examples being Curie temperature distribution and temperature distribution, which are the key properties but ignored previously. To enhance performance, utilizations of higher thermal gradient, magnetically stiffer medium, optimal field etc. have been suggested based upon the theorem. Furthermore, a novel concept, Recording Time Window (RTW), has been proposed. It tightly correlates with performance and serves as a unified optimization standard, summarizing almost all primary parameters. After being validated via spin stand testing, the theorem has been applied to provide solutions for guiding medium design and relaxing the field and heating requirement. This helps solve the issues around writer limit and thermal reliability. Additionally, crosstrack varying field has been proposed to solve the well-known transition curvature issue, which may increase the storage density by 50%.