Adversarial patch attacks are core tools to evaluate adversarial robustness in remote sensing image classification, yet existing methods with fixed patch shapes and single-model optimization yield poor cross-architecture black-box transferability. This paper proposes the Dynamic Patch Shape Selection Framework (DPSF), which integrates a 12-shape patch library, adaptive category matching and multi-model joint optimization with dynamic weight adjustment under a strict 2% patch area constraint. We test DPSF on the UC Merced dataset over eight random split seeds, taking ResNet101, DenseNet161 and VGG19 as held-out black-box models. Under the five-source protocol, DPSF-full achieves mean accuracy drop (${\mathrm{ACC}}_D$) of $0.380 \pm 0.034$, adversarial accuracy (Adv Acc) $0.593 \pm 0.039$, attack success rate (ASR) $0.393 \pm 0.037$ (on clean-correct test samples, averaged over held-out evaluators), union success rate (USR) $0.581 \pm 0.046$, and intersection success rate (ISR) $0.188 \pm 0.023$. With the ResNet50 single-source setting, DPSF-single obtains mean ${\mathrm{ACC}}_D = 0.275 \pm 0.109$, surpassing Sparse-RS ($0.033 \pm 0.012$) and Cross-shaped adversarial patch attack (CSPA) ($0.232 \pm 0.029$). Supplementary results on CIFAR-10, Tiny-ImageNet and Mini-ImageNet illustrate cross-scale transfer trends. This work builds a reproducible digital-domain robustness benchmark for aerial land-use scene classification.