Abstract：The analysis and prediction of deep formation drillability are important for drilling and exploring engineering. In general，the formation drillability index is calculated based on well acoustic logging after the well being completed or at some stages. Using the well logging data，some methods can be employed to calculate the stratum drillability. For the sake of improving prediction technique，the characteristics of formation drillability time series(DTS) are needed. Based on the study of strata drillability time-series of pilot drilling of Chinese continental scientific drilling(CCSD)，two typical indices were put forward. They are characteristics of time-series geometrical form and intrinsic change tendency. These two indices are expressed with correlation dimension(D2) and Hurst(H) exponent，respectively. Based on the analysis of 32 actual well formation drillability profile curves of different oil fields in China，a conclusion was drawn that most of the curves show fractal characteristics. For a drillability time-series，its correlation dimension(D2) can describe the degree of variety and it can be used to express the curve¢s outside geometrical characteristic. The Hurst exponent can be used to estimate the time-series¢ development tendency. The application results show that D2 and H can describe the main characteristics of a DTS perfectly. With these two characteristics，a drillability prediction model based on intelligent neural network was developed. This model was used to predict the formation drillability of CCSD，and it was proven to be practicable.
