Compared with traditional autonomous underwater gliders (AUGs), hybrid-driven underwater gliders (HUGs) can achieve different motion modes by buoyancy driven system and propeller driven system, which are characterized by low power consumption, long endurance and high manoeuvrability, thus being widely used in various oceanographic monitoring missions. An accurate dynamics model with a series of exact enough hydrodynamic parameters is the basis of control system as well as navigation. The main issue of this paper is how to get the accurate hydrodynamic parameters based on the known dynamic model of HUG. In this paper, an HUG named as Petrel which is developed by Tianjin University is selected as the research object. A new method based on large data statistical analysis which combines computational fluid dynamics (CFD) and parameter identification is proposed for hydrodynamic parameter identification. Firstly, dynamic model of Petrel with hydrodynamic model is established. Secondly, lift coefficient is solved by CFD method. Then, other hydrodynamic parameters in the buoyancy driven gliding mode are calculated by large data statistical analysis using a large amount of experimental data. Finally a set of hydrodynamic parameters including the propeller thrust coefficient is obtained. The simulation results show a good agreement with the experimental results, thus verifying validity and availability of this method. And this work paves the way for further design.