Abstract: Metallic magnetic calorimeters are a class of low-temperature particle detectors based on calorimetry, utilizing metallic paramagnetic temperature sensors to convert the temperature rise of an absorber upon the absorption of incident particle energy into a change in magnetic flux, which detected by a superconducting quantum interference device (SQUID). This process enables high-resolution measurements of radiation with exceptional precision. However, metallic magnetic calorimeter signals are inherently weak and highly susceptible to noise, making the extraction of pulse amplitude a significant challenge. Digital signal processing (DSP) plays a crucial role in the data processing of metallic magnetic calorimeter by employing digital filtering techniques to shape the signals and enhance energy resolution. This paper introduces a data acquisition system for metallic magnetic calorimeter based on a field-programmable gate array (FPGA), designed for high-speed signal acquisition, real-time control, signal processing, and display of spectra through an upper computer. Testing of the system yielded successful acquisition of both Co-60 spectra data and noise spectrum of metallic magnetic calorimeter. The linear correlation coefficient of R² reached 0.9781, demonstrates the system’s effective signal acquisition capability. In offline mode, the shaping effects and energy resolution enhancement of traditional nuclear detector signals, simulated, and actual metallic magnetic calorimeter signals were compared using three filtering methods: trapezoidal, cusp, and gaussian filtering. The cusp filter proved to be the most effective, significantly improving the energy resolution of high-purity germanium spectra from 0.598 keV to 0.536 keV. This enhancement validates the correctness of digital signal processing and provides a reliable basis for the selection of digital filtering methods for metallic magnetic calorimeter. In summary, our research presents a comprehensive approach to metallic magnetic calorimeter signal acquisition and processing, highlighting the potential of digital signal processing to overcome the problem of the extraction of weak signals submerged in strong noise. The FPGA-based data acquisition system developed in this study offers an efficient solution for high-resolution radiation measurements.