To solve the problem that the barometer is susceptible to ambient temperature, a pedestrian indoor height localization algorithm based on the fusion of MEMS accelerometer and barometer is proposed in this paper. Firstly, the state of going upstairs and downstairs are separately determined by the peak characteristics of xaxis accelerometer and the interquartile range of the zaxis accelerometer. Then the pedestrian height calculation is carried out by using the barometer data that has been filtered out the air pressure jump value. Finally, when the pedestrian walks horizontally, the calculated height is corrected to an integer multiple of half of a floor height to reduce the height error. The results of experiments show that the pedestrian indoor height localization algorithm proposed in this paper is not easily affected by ambient temperature. The error between the calculated height and the actual floor height is less than 2 m, and the pedestrian can be positioned to the correct floor.