A new genetic triangulation algorithm used for the modeling of 3D data field is presented in this paper.

First, 3D data field was pretreated to a unitary and integral field, and the method of 2D image segmentation was extended and amended, which makes the method be used to segment 3D data field.

Some improvements for the basic ray casting algorithm are made, which make use of various kinds of coherences of the image space and the object space, to accelerate the visualization of the 3D data field.

These data form 3D data sets.

In this paper,we present a multi surfaces rendering method based on volume rendering for 3D data sets.

The visualization of 3D data sets is a multi-disciplinary subject.

Thus ray intersection algorithm is optimized, by which complicated operations of traditional intersection of light ray and parallelogram are replaced, the display speed of 3D data field of the algorithm is raised by using the coherences of object and image space.

The visualization of 3D data not only become more important in the experiments for structures and data processing, but improve the post-processing of FEM software, especially extracting and effectively displaying data.

This paper analyzes the need of spatial analysis in 3D data set and common nest methods,including simple variogram nesting method,equivalent elliptic variogram nesting method and so on,and compares the difference and applicability of them.

A simulation system of down-hole production using the desktop operate system is obtained, which utilizes the 3D modeling software "MultiGen"and the 3D scenes managing software "OpenGVS".