<p>The rapid growth of e-commerce has increased demands on logistics and delivery services. However, logistics enterprises in rural areas face significant challenges, including underdeveloped infrastructure, dispersed populations, and fluctuating demand, all of which contribute to elevated delivery costs. In response, governments have implemented policies and subsidies to encourage logistics enterprises to expand their market presence, improve delivery efficiency, and extend coverage, thereby fostering economic integration between urban and rural areas. Accordingly, we propose a last-mile delivery routing model for e-commerce logistics tailored to rural settings. This model accounts for factors such as fuzzy demand, simultaneous pickup and delivery, government subsidies, and time window constraints. To solve the model, we present an improved Nutcracker Optimization Algorithm (NOA) that integrates fuzzy simulation and solution perturbation to effectively address the problem. The performance of the proposed algorithm is demonstrated through comparisons with other algorithms. Additionally, we conducted a sensitivity analysis to examine the effects of government subsidies and risk preferences on enterprise profits and costs. The findings suggest that enterprises adopting risk-neutral strategies can enhance their distribution profits, while the government can optimize fiscal resource allocation through a combination of two subsidy measures. </p>