Abstract: Reasonable parameter estimation and precision evaluation not only need reliable function model but also need right stochastic model. The influence of stochastic model on the coordinate of Xi'an mobile satellite laser ranging (SLR) station determination is researched in the ways of weight function and outlying observations edit. The coordinate of Xi'an mobile SLR station is determined by using the data of Lageos-1 of all SLR observation stations in the world, which observed from May 1 to June 29, 2016. When the coordinates of Xi'an mobile SLR station are computed, SLRF2014 station positions and velocities are employed. The coordinates of Xi'an mobile SLR station are computed with the orbit of satellite Lageos-1 determination. Three thirty-day orbital arcs and four computation plans are used to compute the coordinates of Xi'an mobile SLR station. There are sufficient numbers of SLR observations to compute reliable orbits, and the values of post-fit range residual root mean square (RMS) is about 2 cm. The difference of the results of the three thirty-day orbital arcs and the four computation plans are compared. By comparison and analysis, we find that the precision of Xi'an mobile SLR observation and the precision of the coordinate computed both reached the level of centimeter. Furthermore, the stochastic model has direct effects on the coordinate of SLR station determination result and reliability. The plan of robust variance component estimation can provide the most precision coordinate, the most stable result, the smallest weighted root mean square (WRMS) and the largest number of SLR data used, when using same computation arc. The more SLR data is used, the higher precision of coordinate is gotten, when using same computation plan.