This paper proposes a two-layer bidirectional gated recurrent unit (BiGRU) with cross-attention mechanism (CA) for the airborne target intent recognition model (BiGRU-CA). First, the air target tactical intent problem is analyzed and described. Then, a two-layer BiGRU network model is constructed to identify the target tactical intent. Subsequently, a cross-attention mechanism (CA) based on the self-attention mechanism (SA) and global attention mechanism (GA) is proposed, which focuses on the key parts of state characteristics and temporal information and satisfies certain interpretability. Comparative experimental analyses with other traditional air target intent recognition models show that the model effectively improves the recognition of the tactical intent of air targets.