Under the influence of the heterogeneous combustion dynamics of coal powder in coal-fired boilers, the combustion products are nonlinearly coupled with the flame radiation absorption coefficient, presenting a chaotic time series, resulting in poor monitoring performance of coal-fired boilers in both time and frequency domains. Therefore, a coal-fired boiler flame temperature field time-frequency characteristic analysis and carbon emission monitoring technology is designed. Based on the principle of high-dimensional continuous spectrum detection of coal-fired boilers under the superposition of medium units, the flame radiation is decomposed into monochromatic light, the radiation intensity at each wavelength is recorded, and the high-dimensional continuous spectrum data of coal-fired boilers is captured at each position; Using the blackbody radiation law and spectral inversion mechanism, the synergistic inversion of coal-fired boiler flame radiation transmission is divided into two forms: concentration field inversion and temperature field inversion and solution, from both the time and frequency domains of spectral data. The chaotic time series is converted into time-frequency series; By aligning the time-frequency values of concentration and temperature fields, the monitoring parameters such as cumulative carbon emissions and instantaneous carbon emission rates can be calculated and mapped to the equivalent cloud map to monitor carbon emissions. The experimental results show that after the application of this method, the spectral fitting degree is 0.998, which meets the requirements of time-frequency characteristic analysis and monitoring technology; The deviation of flame radiation chaos information and carbon emission concentration monitoring were reduced by 2.21 ℃ and 0.93mg/m 3, respectively. The time-frequency analysis results of the flame temperature field of coal-fired boilers are accurate and effective, with better monitoring performance.