The number density and the statistic analysis of defects, cavities and bubbles are important structural parameters in the characterization of the radiation resistant properties of nuclear materials. Transmission electron microscope (TEM) is a powerful tool for the quantitative characterization of irradiation defects. The accuracy of the quantitative characterization is based on the precise measurement of the thickness for the micro-region. The traditional thickness measurement method based on convergent beam electron diffraction (CBED) with tow-beam approximation proposed by Kelly and Allen is a classic technique. But, this technique can not directly measure the thickness of thinner sample with thickness less than 0.6 ξ g ( ξ g is the extinction distance of the diffraction beam g ⇀ ). It is limited in the experiment and post-data process for the thickness measurement of thinner samples. However, for the irradiated materials with high density of cavities or defects, the thinner sample preparation is necessary for the clear observation and precise statistic analysis of the high-density cavities or defects. Therefore, it is urgently needed to develop the TEM-based technology for the thickness measurement of thinner samples. In this work, we use irradiated 316L austenitic stainless steels as research sample. High-density cavities are formed in 316L austenitic stainless steels implanted by high-dose He and then irradiated with high-energy Fe13+ ions. Using focused ion beam (FIB) technique, TEM sample is prepared from the irradiated 316L austenitic stainless steels. To measure the thickness for the thinner micro-region of the 316L stainless steel TEM sample, we propose a new approach based on the CBED technology proposed by Kelly and Allen. Firstly, the extinction distance of specific diffraction beam is evaluated in the thicker micro-region by using Kelly and Allen methodology, and then this parameter is used to calculate the thickness of thinner micro-region of the same sample in which the CBED is recorded under the orientation as that of the thicker region. Using this approach, the density of the cavities and the irradiation swelling ratio in a thinner micro-region of the sample of 316L stainless steels obtained through FIB are quantitatively characterized by TEM. The experimental results confirm that our proposed CBED-based approach is suitable for the thickness measurement of the thinner 316L stainless steel samples with thickness about from 20 to 50 nm. Furthermore, to present the advantage of CBED-based method for thinner-region thickness measurement, we compare it with the thickness measurement method based on electron energy loss spectroscopy (EELS). The thickness measurement results by using EELS are the total thickness of the TEM samples, including the thickness of the surface amorphous layer introduced through the sample preparation. In contrast, the CBED-based method can provide the intrinsic thickness of the crystalline samples without the influence of surface amorphous layer. Finally, we discuss the experimental factors which influence the thickness measurement results of thinner samples by using our proposed experimental approach and the related optimization methods. It is envisaged that our proposed approach can be extended to the micro-region quantitative statistic analysis of different type of defects, such as dislocation loops, cavities, bubbles and stacking fault tetrahedrons in ion irradiated crystalline materials.