In order to investigate the relationship between variations in carbon dioxide （CO₂） concentration and seed vitality during respiration， a CO₂ photoacoustic sensing system utilizing a differential Helmholtz photoacoustic cell was designed. The system's performance was evaluated by measuring the CO₂ production rate of south japonica rice seeds exhibiting varying levels of vitality. An independent signal processing module based on a field programmable gate array （FPGA） development board was developed to facilitate precise regulation of laser drive current and demodulation of photoacoustic signals. Through optimization of modulation frequency， laser modulation depth， working pressure， and other system parameters， optimal performance was achieved， with an integration time of 1 089 s and a detection limit of 0.6×10^-6. In accordance with national standard GB/T 3543.1， a germination experiment involving 30 g of south japonica rice seeds was conducted. Linear fitting analysis was performed using CO₂ evolution concentration data obtained from the photoacoustic sensing experiment. The results demonstrate that the seed respiration rate of south japonica rice is directly proportional to seed vitality， with a correlation coefficient of 0.979. This finding provides a valuable reference for assessing rice seed vitality through photoacoustic spectroscopy.