Research for detecting or obtaining radionuclide by gamma energy spectrum data acquisition and process system is one of the key issues about intelligent measurement of gamma-ray spectrum. For this reason, a software and hardware implementation schematic design based on ARM ( Advanced RISC Machines) + DSP ( Digital Signal Processor) architecture for gamma energy spectrum data acquisition and processing system is proposed. The paper discusses in detail some key technologies such as communication interface design between microcontroller ARM and digital signal processor DSP,distribution scheduling under multi-task in the ARM-Linux,DSP handling procedures for multi-channel A / D high-speed sample. At the same time,because the traditional Gaussian fitting to determine the boundary of peak is not ideal,it puts forward a weighting factor of Gaussian function least squares fitting realize boundary determined. Finally gamma-spectrum data from sodium iodide NaI( TI) scintillation detector is tested and processed in the new system. The results show that gamma energy spectrum data acquisition and process system is perfect functionality, stable and convergence in unimodal. Compared with data from conventional energy spectrometers,the system can keep better energy resolution in a wide range of pulse pass rate. Research for detecting or obtaining radionuclide by gamma energy spectrum data acquisition and process system is one of the key issues about intelligent measurement of gamma-ray spectrum. For this reason, a software and hardware implementation schematic based on ARM (Advanced RISC Machines) + DSP (Digital Signal Processor) architecture for gamma energy spectrum data acquisition and processing system is proposed. The paper discusses in detail some key technologies such as communication interface design between microcontroller ARM and digital signal processor DSP, distribution scheduling under multi-task in the ARM -Linux, DSP handling procedures for multi-channel A / D high-speed sample. At the same time, because the traditional Gaussian fitting to determine the boundary of peak is not ideal, it puts forward a weighting factor of Gaussian function least squares fitting realize gamma-spectrum data from sodium iodide NaI (TI) scintillation detector is test ed and processed in the new system. The results show that gamma energy spectrum data acquisition and process system is perfect functionality, stable and convergence in unimodal. Compared with data from conventional energy spectrometers, the system can keep better energy resolution in a wide range of pulse pass rate.