Abstract: For the problem of optimizing interference utility concerning the parameters of noise convolution kernels in clever noise convolutional jamming, combined with the characteristics of multi-pulse coherent processing in multi-pulse radar systems, this study establishes a mathematical model of multi-pulse smart noise convolutional jamming based on intermittent sampling. It explores the mapping relationship between convolution kernel length and interference effectiveness in convolutional modulation. Theoretical analysis is validated through corresponding simulation experiments. The results demonstrate that the interference signal in multi-pulse radar systems is derived from coherent accumulation of single pulses, where the convolution kernel length exhibits a square-root growth relationship with interference signal amplitude, an inverse proportional relationship with effective bandwidth, and a positive correlation with spectral concentration, thereby confirming the theoretical analysis. This work provides valuable references for enhancing radar interference effectiveness in complex electromagnetic environments.