Underground coal gasification (UCG) is an environmentally friendly recovery process for coal, especially for deep coal resources. As a potential technique, it will become an important method for the effective development of coal resources in such a time of energy transition. Currently, how to accurately numerically simulate the UCG process with respect to gasification chamber evolution and recovery performance at a field scale is still unclear. For the UCG process, a method of Linear Controlled Retracting Injection Point (L-CRIP) is the most effective well pattern. In this paper, a large-scale numerical simulation model is developed using porous medium approach to simulate the L-CRIP process in a deep coal seam. First, to get a better understanding of the chemical process in UCG, a series of thermogravimetry (TG) tests are conducted. Then, nine chemical reactions including coal pyrolysis, combustion, and gasification are considered in our model. The reaction frequencies are described by an Arrhenius correlation. The coal permeability is simulated by an exponential function of fluid porosity which is calculated by the mass conservation law and chemical reactions. Furthermore, using this model, the evolution process of the gasification chamber during a UCG process is investigated. Simultaneously, a series of simulation runs are performed to study the effect of sensitive factors (e.g., oxygen injection rate, steam/oxygen ratio and retraction rate.) on flow rate, composition, and heating value of the produced syngas. Results indicate that the L-CRIP process can effectively develop deep coal resources. For the simulation case discussed in this paper, the maximum syngas daily production rate can reach above 1.0×104 m3/d. Based on the profiles of temperature, porosity, and char concentration, it is shown that the shape of the gasification chamber quickly expands to a sphere in the early stage, while subsequently it gradually to like a teardrop, possibly controlled by the retraction distance, the retraction rate, the pressure drop, and the permeability of the domain. The optimal steam/oxygen ratio and retraction rate is 2:1 and 0.5 m/d, respectively. Besides, the operation strategy of variable oxygen injection rate, variable steam/oxygen ratio, and variable retraction rate within an entire UCG process is highly recommended. This paper provides a practical framework for the simulation of the UCG process at a field scale. It sheds some important insights for gasification chamber evolution and recovery performance of UCG process using L-CRIP technique in a deep coal seam.