The signal acquisition and processing ability of distributed temperature sensors based on Raman scattering (DTS-Raman) directly determines the accuracy, spatial resolution, temperature resolution, and other key indicators of the entire system. Real-time and effective processing is the key to ensuring the practicability of the field system. Owing to the long distance of the distributed temperature sensors (DTS) system, the energy of the weak Raman scattering is further reduced when it reaches the avalanche photodiode (APD) detector because of fiber loss, which results in a decrease in the signal-to-noise ratio (SNR) and an increase in the temperature measurement error. To improve the SNR and ensure the accuracy of the demodulation temperature, the signal acquisition and processing module must be able to perform real-time and fast processing of large amounts of data. The existing processing structure based on the acquisition board requires a special personal computer (PC) and host processing software, which limits the application of DTS system. Therefore, a real-time acquisition and processing scheme based on field-programmable gate array (FPGA) is proposed in this paper. A set of cyclic shift register sequence (CSRS) with Ping-Pang functions is designed to improve the storage efficiency. According to the period of the laser source, the accumulated results stored in the CSRS and echo pulse data are accumulated and shifted in real time to realize circular utilization of storage resources. A DTS-Raman system prototype and test platform are built to verify the effects of the proposed scheme. The experimental results show that the scheme can realize a temperature accuracy within ±0.475°C in real time and has a high resource utilization efficiency.