Abstract: Focusing on Guizhou Province, remote sensing data from Himawari–9 AHI and FY–3D MERSI were collected. Four classification methods of Maximum Likelihood, Neural Network, Support Vector Machine, and Random Forest were used to extract smoke information, facilitating fire point detection and identification through contextual analysis. Additionally, the potential of Himawari–9 AHI and FY–3D MERSI for early wildfire monitoring was evaluated. The results show that the smoke misclassification error of Random Forest is smaller than other methods with the Kappa and Sørensen of 0.79 and 0.76, respectively. The smoke classification results based on two image data show better consistency.The smoke mask image is better overlapped with the satellite imagery while the FY–3D MERSI resulting higher accuracy. The wildfire fire point detection precision of FY–3D MERSI and Himawari–9 AHI based on the contextual fire-point detection reached 89.98% and 79.80%, respectively showing better wildfire detection of FY–3D MERSI. Given FY–3D has fine spatial resolution and Himawari–9 has higher temporal resolution, integrating two satellite data can effectively improve the timeliness of wildfire monitoring, which has certain reference value for the practical work of fire monitoring and early warning and suppression decision making support, etc.