Swarms of mobile wireless-connected sensors are increasingly deployed for applications such as monitoring, surveillance, and safety-critical operations. Quantifying end-to-end (e2e) delay performance guarantees in these scenarios is paramount. In this paper, we present a theoretical approach using Stochastic Network Calculus (SNC) with Moment Generating Functions (MGFs) to characterize e2e delay bounds in Mobile Wireless Sensor Networks (MWSNs). Our study focuses on a network composed of two segments: the first segment includes multiple nodes connected via a contention-based channel using the Distributed Coordination Function (DCF), while the second segment consists of a link prone to disconnections due to the mobility of nodes in the first segment. We model the first segment by calculating the expected per-packet service time in a non-saturated homogeneous contention channel and the second segment using a Discrete Time Markov Chain (DTMC). Initially, we derive a mathematical expression that correlates the offered load with the saturation status of each node’s queue in a non-saturated contention channel with homogeneous nodes. We then provide numerical e2e delay bounds for an illustrative example of a first responder network, quantifying the effects of non-saturated traffic, communication range on the head-sink link, and scheduling algorithms across different network sizes. Finally, we compare the derived e2e delay bounds with network simulations to assess their accuracy and reliability.