Abstract: Low-frequency electromagnetic field simulation is the fundamental and key technology for the artificial intelligence of electrical equipment. Based on the development history of computational electromagnetics since the 1960s, this paper discusses the state of art of several key issues in low-frequency electromagnetic field simulation and looks forward to the current development trend in the field of numerical simulation. Firstly, the basic equations and numerical discretization methods of low-frequency electromagnetic field calculation are introduced. Secondly, the issues including homogenization, anisotropy and hysteresis model in material modelling are analyzed and discussed. Subsequently, the special characteristics of mesh and discrete format in low-frequency electromagnetic field simulation are explained from several aspects, such as multi-scale problem, skin effect, open boundary as well as deformation and motion; the current development trends of electromagnetic-thermal-structural multi-physics coupling are then analyzed; physical model-based and data-driven modelling approaches as well as recent work in digital twin model building are introduced. Finally, the development trend of numerical simulation in optimal design, digital twin, artificial intelligence and new generation of simulation software is analyzed and discussed. At present, multi-scale and multi-physics coupled modeling and simulation is still a focus of research in the field of electromagnetic field computation. The integration of the traditional numerical simulation with emerging computing technologies and artificial intelligence is becoming more and more significant. Under the cutting-edge needs of smart manufacturing and digital twins, electromagnetic field computation faces the challenges of efficient iterative simulation and real-time simulation. Physics and data-driven based surrogate modeling is a key technology to resolve the two issues. Industrial CAE simulation software architecture is undergoing a profound change, there is an urgent need to enhance the core technology and application ecology of China's self-developed CAE simulation software.