In the intelligent robot suit, the output performance of piezoelectric sensor array is easy to be influenced by nontarget factors, such as environment temperature, voltage fluctuation and so on, resulting in obvious measurement precision decrease due to the intercross interference among the sensors. Aiming at the above problem, a new tactile sensor array as well as signal processing method for distributed pressure measurement is presented. The processing flow of sensor signal is designed to extract and effectively process the useful data in the confidence distance matrix and relation matrix by means of mathematical statisticsbased approach. The fusion processing technique with maximum probability method and maximum likelihood method for multisensor information is presented. The experimental result shows that the proposed method can eliminate the temperature impact on pressure sensor to reduce the error influence from uncertainty of single sensor and enhance the stability and measure precision for distributed pressure measurement data, which outstandingly improves the reconstruction precision of pressure image.