Photon counting laser altimetry could detect the continuous elevation of the earth’s and vegetation’s surface. But， whether a single photon event could occur in a certain period is a random event according to echo energy and detection probability. Therefore， the set of received single photon events can not accurately represent the strength of an echo， which makes it difficult to determine the starting position in monopulse detection， and may cause an error when estimating vegetation’s canopy height. By establishing the photon counting detection model of vegetation， the error of canopy height caused by the principle of single photon detection is analyzed and calculated in this paper. Experimental results showed that the average elevation of the first detected signal photons was 2.435m lower than the DSM provided by LIDAR data from taiga forest. Still， it is helpful to reduce the error by increasing the energy or reducing the divergence angle of of emitted laser. And the errors of forests with higher canopy density and leaf area index are smaller accordingly.