Forest fires are major natural disasters worldwide, posing severe threats to the ecological environments, human lives and property. To address forest fires more effectively, this study introduces the concept of digital twins and employs the fire dynamic simulation software PyroSim, coupling five factors: ignition points, slope, wind speed, relative humidity and ambient temperature. A microscopic forest fire scenario in Pinus yunnanensis forests is constructed for refined simulations, aiming to investigate the influence of these factors on forest fire spread and to analyze the dynamics fire development process. A sliced dynamic simulation method is proposed to mitigate the lag issues in simulations under fixed environmental parameters and initial fire source conditions. The research results indicate that the digital twin-driven sliced dynamic simulation can reflect the actual behavioral characteristics of forest fires in a more realistic and efficient manner, thereby providing an important references for the study of forest fire suppression strategies.