Crustal movement will discharge CO₂ and other gases to the surface, and the surface concentration of CO₂ near the fault zone will be abnormal before the earthquake. High-precision measurement of CO₂ gas near the seismic zone can provide important help for the analysis of earthquake precursors. At present, the main methods for measuring CO₂ concentration include non-dispersive infrared analysis technology, electrochemical technology, chromatographic analysis technology, etc. However, the above methods generally have the disadvantages of being easily disturbed by its background gas, low accuracy, and unable to achieve real-time monitoring. Tunable diode laser absorption spectroscopy (TDLAS) technology has the advantages of not being disturbed by its background gas, high accuracy, and real-time monitoring. In recent years, it has become a research hotspot at home and abroad and has been widely used in the field of gas detection. In this paper, a high-precision CO₂ detection system is developed by using tunable diode laser absorption spectroscopy technology. In this paper, a high-precision CO₂ detection system for seismic monitoring is established. The tunable diode laser absorption spectroscopy technology is adopted, and the wave number 4 978.202 cm^-1 is selected as the absorption spectral line of the CO₂ detection system (Fig.1). A multi-channel unit with an effective optical path of 40 m is adopted, and STM32 is used as the control equipment and data processing core equipment (Fig.2). For the detector noise and optical interference fringe noise in the system, Kalman-wavelet analysis algorithm is used to filter and improve the system. The system uses Kalman-wavelet analysis method to eliminate the influence of detector noise and optical fringe interference. The experiment shows that the second harmonic signal to noise ratio of the system at 50 ppmv CO₂ concentration is 2.06 times higher than that before filtering (Fig.3). Under different CO₂ concentrations (50 ppmv, 300 ppmv, 1 000 ppmv, 4 000 ppmv, 8 000 ppmv), the system error is 2.57%-2.66% (Fig.4). When the system measures CO₂ at 4 000 ppmv concentration, the detection precision reaches 20.9 ppmv (Fig.5). According to Allan variance analysis, the method detection limit (MDL) corresponding to the integration time of about 61s is 5.2 ppmv (Fig.6), which realizes the high-precision measurement of CO₂ gas. This paper develops a high-precision CO₂ detection system for seismic monitoring. The system adjusts the current injected into the DFB laser to make its output central wavelength at 2 008 nm and serve as the detection light source of CO₂. In order to improve the lower detection limit of CO₂ gas concentration, the system uses a self-developed cylindrical mirror multi-pass cell with an effective optical path of 40 m. The multi-pass cell can work stably in the temperature range of 0-40 ℃ and the pressure range of 1.333-101.325 kPa to ensure the reliability of the system in the field measurement process. The system control TEC realizes the temperature control of the controlled object, and the control precision of the temperature control system in the laboratory can reach 0.01 ℃. The Kalman-wavelet analysis algorithm is used to filter the system noise, and the frequency of optical fringe interference in the frequency domain is similar to that of cosine wave in the time domain, so as to separate it and remove the optical fringe interference. The experimental results show that the accuracy, precision and the method detection limit of the system are improved after filtering. The system combined with this method can make the geochemical gas measurement have a broader application prospect and provide important help for the accurate analysis of earthquake precursors.