Due to lack of experimental data of the inner shell ionization cross sections induced by low-energy positron, advanced theoretical models developed in recent years cannot be correctly evaluated, and the application of slow positron beam technique is greatly limited. Here we present the method of obtaining reliable experiment data of atomic inner-shell ionization cross section by positron impact. In this work, the slow positron beam device is used to generate 8-9.5 keV positron beams impacting on a pure thick Ti target, and the silicon drift detector (SDD) is adopted to collect the X-ray spectra produced by positrons impacting on thick Ti target, and the incident positron numbers are obtained by applying an HPGe detector to on-line collect annihilation photons. Then the experimental characteristic X-ray yields of Ti K shell impacted by 8-9.5 keV positron could be acquired. Meantime, the simulated characteristic X-ray yields are acquired by the PENELOPE program simulating the experiments. In the comparison between the experimental yields and the simulated yields based on two sets of different inner shell ionization cross section database in the PENELOPE code, i.e. the optical data model (ODM) and the distorted-wave Born approximation model (DWBA), there is a large difference between the simulated data from the ODM theoretical model and the experimental values, while the simulated yields from the DWBA theoretical model are in good agreement with the experimental results. Accordingly, a correction factor is introduced to modify the DWBA theoretical model database which is used in the PENELOPE, and then the experimental process is re-simulated. When the simulated yields and the experimental yields are in the highest consistence, the reliable Ti K shell ionization cross sections impacted by 8-9.5 keV positron could be obtained. The biggest advantage of using this method to obtain atomic inner-shell ionization cross section impacted by positron is that the effects of the multiple scattering of incident positrons in the thick target, from the bremsstrahlung and annihilation photons, and other secondary particles on the experimental characteristic X-rays do not need calculating (the calculation method that has been developed previously cannot give the more correct result about the contribution of the multiple scattering of incident positrons, from the bremsstrahlung and annihilation photons, and other secondary particles to characteristic X-rays).