Abstract: Human visual sense is the main way to get information. Only the lifelike and unartificial virtual vision can make the users immersed in a virtual environment. The virtual vision of DVENET generated from real terrain includes realistic culture features and natural phenomenon. Here introduced are some new methods during the generation of DVENETs virtual vision, such as environment geometric modeling, the combination of geometric model, special effect rendering, cutting apart and dispatching method, and vision coinciding.