In this paper, the Monte Carlo (MC) method was used to investigate the optimization of pyrometer design and the uncertainty assessment for shock-induced temperature measurements. The statistic T_90%, which was defined by p{｜T-T_true｜>T_90%}=0.1, was calculated for 10 kinds of designs, and each case was simulated 2×10⁵ times to make the statistical sample. Four conclusions are presented: (1) Using more channels in a pyrometer can surely, but slightly improves the measurement accuracy of shock-induced temperature. (2) In the most case, measurement accuracy evidently depends on the accuracy of each channel. Generally, when the channel uncertainty is reduced from 10% to 5%, the experimental accuracy can be increased by over 150%. (3) The measurement accuracy also obviously depends on the channel's wavelength range. (4) The measurement accuracy can be largely improved by using those wavelength channels at which the difference of the radiation energy measured is as large as possible.