Drones in emergency delivery operations face complex electromagnetic interference, including active jamming such as suppression and deception types, and passive jamming like chaff, which seriously threatens the reliability of radar detection. Constrained by platform payload and computational power, traditional anti-jamming techniques are difficult to apply directly. This study proposes a set of optimization strategies centered on lightweight design and adaptability, including interference-aware waveform optimization, joint spatiotemporal processing, jamming evasion methods, and lightweight algorithm design. Through theoretical analysis and numerical simulation, the strategy achieves a Signal-to-Interference-plus-Noise Ratio (SINR) improvement of up to 22 dB under compound interference, maintains a stable target detection probability above 90%, and reduces computational complexity to only 30–45% of that of traditional methods. This effectively balances performance with resource constraints, thereby enhancing the reliability of delivery missions.