In order to improve contrast and resolution in photoacoustic tomography (PAT), the filtered photoacoustic (PA) signal of tissue phantom was often deconvoluted by transducer impulse response to restore its spectrum. Because the cut-off frequency of band-pass filter was determined manually, it couldn’t suppress noise effectively, it was hard to get static deconvolution result. To solve this problem, a Wiener filter deconvolution algorithm was presented. The transducer impulse response was measured by a point photoacoustic source. The noise was suppressed by Wiener filter during deconvolution. The filter’s parameter was estimated by discrete wavelet transform (DWT). The photoacoustic image was reconstructed by time domain back projection algorithm. Both numerical simulation and experimental results demonstrated that the noise was suppressed effectively during deconvolution. The imaging contrast and resolution was improved.