Most of infrared image denoising methods focus only on the space domain or the frequency domain. Aiming at the shortcoming, an infrared image denoising algorithm is proposed based on wavelet transform threshold shrinkage method. The algorithm has the characteristic of analyzing time domain and space domain simultaneously, and wavelet transform with a concentration characteristic to the determined signal is used to make the power of a signal on a small number of coefficients in wavelet transform domain. And then, wavelet coefficient is threshold, the low amplitude noise and unexpected image information in wavelet transform domain can be removed. Experimental results show that image quality can be improved effectively by threshold shrinkage denosing algorithm based on wavelet transform. The visual effect of the image is enhanced and the expected image information is obtained accurately.