Abstract: Thermal neutron detection plays an important role in materials science, nuclear energy, and medical applications. Compared with other mainstream thermal neutron detectors, ionization chambers offer advantages such as simple structure, low cost, good radiation resistance, and excellent energy resolution. The optimization study of grid ionization chamber for thermal neutron detection based on the ⁶Li(n, t) reaction was presented. The Geant4 simulation toolkit was employed to investigate the effects of the ⁶LiF conversion layer thickness, working gas type (90%Ar+10%CO₂ versus CF₄), and ionization zone thickness on the energy spectrum, emission angle, and detection efficiency of secondary particles. When the thickness of the ⁶LiF coating is approximately 3 μm and 20 μm, respectively, the optimal detection efficiency is achieved for α particles and tritons. Considering both the preparation process and detection performance, a ⁶LiF coating with a thickness of about 1 μm ensures a uniform and dense coating while achieving relatively good detection efficiency. Compared with the Ar/CO₂ mixture, CF₄ exhibits higher stopping power. With an ionization zone thickness of 12 mm, the detection efficiency for tritons reaches saturation, whereas the Ar/CO₂ mixture requires 35 mm to achieve comparable performance, thereby significantly reducing the sensitive volume of the detector. Experimental validation was conducted using a neutron irradiator at a micro-reactor facility. The experimental results are in good agreement with the simulation data. This study provides a design basis and parameter optimization framework for the engineering application of high-performance and low-cost thermal neutron detectors.