Objective: Densified wood (DW) is a novel functional material that has garnered significant attention in recent years owing to its exceptional mechanical properties. In comparison with natural wood (NW), DW exhibits a more compact structure, which augments its strength and confers enhanced fire resistance. These advantages position DW as a promising candidate for utilization as a structural material in mid-and high-rise timber buildings, where strength and fire safety are paramount. Nevertheless, despite its potential, the current understanding of the flammability of DW and its fire spread characteristics under real fire scenarios remains limited; this dearth of knowledge imposes substantial constraints on its large-scale implementation in the construction sector. However, a paucity of research has emerged on the coupled effects of wood density and thickness on fire spread in DW. Concurrently, the regulatory influence of external thermal radiation on the fire spread characteristics of DW remains to be fully investigated. Addressing these gaps is imperative for developing a comprehensive theoretical foundation for the safe and reliable use of DW in modern building design. Methods: This study integrates theoretical analysis with systematic experimental testing to explore these issues. The fire spread behavior of DW was investigated under an external radiant heat flux of 8 kW/m². The specimens' grain direction was maintained parallel to the anticipated fire spread direction, and samples were meticulously prepared to ensure the absence of visible knots, thereby mitigating potential irregular burning effects. Each specimen measured 600 mm in length and 30 mm in width, with two distinct thicknesses: 2 mm and 8 mm. DW specimens with varying density gradients (281.53-973.61 kg/m³) were obtained by compressing NW of different initial thicknesses to these unified final thicknesses. A multiparameter fire spread testing platform was designed and constructed, consisting of three main components: a radiant heating system, specimen support system, and multiparameter data acquisition system. The initiation of the combustion process was facilitated by the application of a linear butane flame to the surface of the DW samples, thereby ensuring uniform ignition conditions across all experimental iterations. During the experiments, flame morphology, solid-phase temperature, and gas-phase temperature were recorded in real time. The analysis of flame images was conducted using digital image processing techniques to extract key parameters, such as the flame front position and flame height, thereby facilitating qualitative and quantitative assessment of combustion behavior. Results: Distinct fire spread behaviors were observed to depend on specimen thickness and density. When the specimen thickness was minimal, the fire spread rate initially increased and subsequently decreased with increasing density, and a critical turning density of 573 kg/m³ was identified. Conversely, as specimen thickness increased, the fire spread rate exhibited a monotonic decrease with increasing density. Flame height exhibited a variation trend that was consistent with the fire spread rate. Preliminary theoretical analyses indicated a positive correlation between wood density and the release of combustible gases during the combustion process, which in turn resulted in taller flames. However, with further increases in density and thickness, a decrease in flame height was observed owing to the growth of the material's thermal inertia, which inhibited rapid heat transfer. Furthermore, wood density was found to influence the rate of fire spread by affecting gas-phase and solid-phase heat transfer mechanisms. In the case of thin specimens, the initial stage of fire propagation was predominantly characterized by gas-phase heat transfer, while as density increased, the mechanism transitioned toward a combined effect of gas and solid phases. In the case of thick specimens, solid-phase heat transfer emerged as the predominant factor influencing the propagation of fire. Thermal thickness was employed as a classification criterion for the specimens, in accordance with heat transfer theory principles. Theoretical fire spread rates were subsequently calculated. The theoretical values exhibited a strong correlation with the experimental results, as evidenced by a correlation coefficient of R²=0.88, thereby substantiating the model's predictive capacity. Conclusions: The present study offers a comprehensive understanding of the fire spread behavior and heat transfer mechanisms of DW under external radiation. The findings reveal the complex interplay between density, thickness, and heat transfer mode; they also highlight the critical role of external radiation in shaping fire dynamics. This research contributes to the advancement of knowledge in the field by enriching the theoretical framework of fire spread and offering valuable guidance for the safe design and application of DW in modern construction. The findings of this study can assist in overcoming the current barriers to the large-scale adoption of DW, thereby promoting its use as a sustainable, high-performance, and fire-safe material. This, in turn, will contribute to the green development of the construction industry.