An effort towards understanding of the stray light problems for the Xinglong 2.16-m telescope was presented to estimate the stray light performance of the telescope itself and provide a method for improving the stray light suppression. The stray light analysis for 2.16-m telescope model, which consists the onion shaped dome, telescope structure, equatorial mount and telescope optics, has been performed with two cases (1) point to 60° and (2) point to zenith, in both azimuth and elevation. A simplified mechanical model obtained from the detailed telescope system geometry was used to increase the simulation efficiency. The source setting, property assumptions and ray tracing methodology for stray light simulation are described. The critical and illuminated objects were identified to understand the stray light propagation path in system. The Point source normalized irradiance transmittance (PSNIT), which is generally used for assessing stray light and uncorrelated to entrance aperture was calculated utilizing the stray light software TracePro to evaluate the stray light performance of the 2.16-m telescope with a series of off-axis angles. It shows that the PSNIT values are less than 10 -7 when off-axis angles are larger than ±20°. The dominant objects for stray light (primary and secondary mirror, telescope structure and dome) were identified to give advice for performance improvement. The analyses indicate that significant benefit can be realized with adding only 5 vanes inside the bottom portion of the secondary baffle. In the case of point to zenith, the PSNIT values will decrease about 40% at average.