When there are scanty core data and more log data,feedforward neural network is a very effective method to recognize the logging lithology.

The numerical well logging-microfacies model is established by means of extracting seven key logging facies elements from several well logging parameters which can reflect the variation of sedimentary microfacies and extracting several synthesis parameters by principal constituent analysis.

Logging data contain abundant information on geology,especially,sedimentary environment from which various parameters characteristic of different sedimentary environments can be obtained in order to establish models for sedimentary facies and logging facies in the studied area.

With the use of Logging information such as natural potential,natural gamma,imaging logging,and diplog,the paper establish the correspondence between the sign of sedimentary facies and logging facies,then use the method of Core logging calibration methods divising the type of single well facies and microfacies.

A lot of new knowledge such as sedi- mentation geological characeristics and electrofacies rules,has also been gained from systematically theoretical analysis,Which lays solid foundation for further development in Sulige gas field.

Gamma ray curve of the Upper Paleozoic in Daniudi gas field can be divided into 4 basic electrofacies types, including bell - shaped, funnel-shaped, box-like and finger - like electrofacies.

This paper introduces the models of electrofacies identification, based on the welllogging information, the model can be used to quantitatively fix the space location of electrofacies in drilled formation, and to set up electrofacies.