Reducing latency is essential for the performance of real-time applications like autonomous driving, smart cities, and healthcare monitoring within the edge-cloud collaborative computing framework. Traditional resource allocation techniques often struggle to adapt to dynamic network conditions and fluctuating computational needs, resulting in higher latency and decreased efficiency. This study introduces an innovative approach to dynamic latency reduction through a hybrid algorithm that integrates the Flower Pollination Algorithm (FPA) with the Backward Induction Algorithm (BIA). Our proposed method tackles resource allocation challenges by utilizing FPA’s global search abilities and BIA’s future state assessments. We mathematically formulate the latency minimization problem, considering constraints specific to edge and cloud resources. The hybrid algorithm allocates resources dynamically, ensuring optimal performance while adjusting to varying network conditions. Extensive simulations and comparative analyses reveal that our approach significantly lowers latency compared to current methods. The findings emphasize the potential of the proposed framework to improve the efficiency and reliability of edge-cloud collaborative computing environments. This research offers a scalable and adaptive solution for resource allocation, advancing performance in various real-time applications.