In the sound event localization and detection (SELD) task, Transformer-based models have demonstrated impressive capabilities. However, the quadratic complexity of the Transformer’s self-attention mechanism results in computational inefficiencies. In this paper, we propose a network named SELD-Mamba, which utilizes Mamba, a selective state space model, to capture long-range contextual information while maintaining computational efficiency. The SELD-Mamba is built upon the Event-Independent Network V2 (EINV2) and employs bidirectional Mamba (BMamba) blocks instead of Conformer blocks. In addition, we implement a two-stage training method to balance the performance of different tasks, with the first stage focusing on sound event detection (SED) and direction of arrival (DoA) estimation losses, and the second stage reintroducing the source distance estimation (SDE) loss. The experimental results on the 2024 DCASE Challenge Task3 dataset show the superior performance of SELD-Mamba, which demonstrates the great potential of Mamba in SELD.