Microseismic detection events are critical in monitoring subsurface activities, including hydraulic fracturing, enhanced oil recovery, carbon dioxide, or natural gas geological storage and reservoir characterization, to guarantee safe and efficient energy extraction. This presents significant challenges as it generates large amounts of data. Despite the recent machine learning techniques being increasingly integrated into fiber-optic distributed acoustic sensor (DAS) systems to enhance their intelligent recognition capabilities, there is still a need to solve this. In some research, it has been observed that computational speed is time-consuming and overfitting, thus necessitating a more extensive investigation of this analysis. This study proposes a novel approach using DAS data to enhance microseismic event detection's precision, overfitting issue, and interpretability. This approach utilizes a specifically designed neural network architecture. The deep learning approach is highly effective for the real-time management of the substantial amounts of data recorded by DAS equipment. Three phases of research methodology are proposed. The contribution of this research is that spiking neural network architecture for microseismic detection will bring advancements in microseismic monitoring. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.