The invention provides a high-siliceous shale longitudinal dessert prediction method which comprises the following steps: S1, collecting a shale
core sample, and obtaining actual measurement experiment data and
logging data of shale; wherein the actual measurement experiment data of the shale comprises actual measurement TOC, nuclear magnetic
porosity and siliceous mineral relative content; s2, establishing a relationship among the actually measured TOC, the nuclear magnetic
porosity and the relative content of the siliceous minerals, and determining the
lower limit values of the relative content of the siliceous minerals, the TOC and the nuclear magnetic
porosity of the gas production section; s3, establishing a
logging siliceous mineral relative content calculation model; s4, determining a shale dessert section; and S5, verifying the accuracy of the shale dessert section. The method takes the relative content of siliceous minerals as a key parameter, and is simple and clear; the relative content of the siliceous minerals is also the most important brittle minerals in the shale, the higher the relative content of the siliceous minerals is, the better the
brittleness of the shale is, and fracturing is facilitated, so that the high-siliceous-content section is also an
engineering dessert section, and the method gives consideration to prediction of geological and
engineering double desserts.