Combined with the advantages of Wavelet threshold denoising method and data fusion technology. A data processing method of thermal measurement based on wavelet transform and data fusion is proposed. After decomposing the thermal measurement signal of each measuring point separately by wavelet decomposition, the high frequency components are first filtered by the threshold noise reduction method, and then the signal of each measuring point after noise reduction is got by wavelet reconstructing to reduce the influence of noise, and then , we normalize the signal of each measuring point, and make the data fusion based on the least mean square of the normalized measuring points, so as to obtain a better reconstructed signal. The experimental results show that this method improves the accuracy and reliability of the measured data and makes the signal features more apparent.