Marine weather radars play a crucial role in ensuring maritime navigation safety. However, when operating in open-ocean environments, radar performance is significantly degraded by sea clutter and dense, multi-source electromagnetic interference (EMI) from co-frequency radios, navigation radars, and satellite communication systems. These coupled factors lead to substantial uncertainty in the credibility of conventional clutter-target simulation models.

This study aims to develop and validate a comprehensive credibility assessment framework for marine weather radar electromagnetic simulation models under complex “clutter-interference-target” coupling environments, providing a quantitative basis for model validation and online calibration.

A mixed-distribution model was employed to describe the joint statistical properties of sea clutter and EMI. A tri-domain feature space (time-frequency-space) was constructed, and multidimensional dynamic time warping (MD-DTW) was used to quantify discrepancies between simulated and measured signals. Finally, a Bayesian network integrated statistical and feature-level results to yield system-level credibility with uncertainty bounds.

Monte Carlo simulations with 500 iterations and bootstrap estimation demonstrated that the proposed method achieves an average credibility of 0.82 at the statistical level, 0.88 at the feature level, and 0.86 at the system level. Compared with traditional methods that ignore EMI, the framework improves credibility by approximately 11%.

The proposed three-tier “distribution–feature–system” framework enables full-chain, multi-dimensional quantification of model credibility under sea clutter and EMI coupling. This approach enhances the reliability of radar performance assessment in complex electromagnetic environments and provides a rigorous basis for adaptive calibration and resource management.