In order to solve the current problem of poor portability and low sampling rate of wearable functional near-infrared spectroscopy (fNIRS) detection equipment at home and abroad, a modulation and demodulation method of fNIRS based on FPGA is proposed in this paper. Firstly, the "frequency division -time division multiplexing" modulation of the light source driving signal is realized based on FPGA, and the frequency division is carried out between multiple wavelengths of a single light source, and the time division is carried out between multiple light sources, so as to improve the sampling frequency (frequency division) while taking into account the size of the system volume (time division). Secondly, the embedded digital demodulation and digital filtering of the detection signal is realized based on FPGA, so as to facilitate the interpretation of the blood oxygen signal and embedded application directly based on the embedded system. Compared with the current portable cerebral oximetry acquisition equipment, the design and development of FPGA-based embedded near-infrared spectroscopic imaging system improves the sampling rate, decreases the requirements for FPGAs, and reduces the size to increase portability. PGA-based signal demodulation lays an important foundation for the implementation of embedded applications and is expected to be suitable for more application scenarios.