Accurate node localization is crucial for applications such as environmental monitoring and target tracking in wireless sensor networks (WSNs). Distance Vector-Hop (DV-Hop) is a classical range-free localization algorithm widely used due to its hardware-free implementation. However, its performance degrades significantly in the presence of topological holes, where path distortion leads to inaccurate hop count estimation. To address this challenge, this paper proposes IZOA-DV-Hop, an enhanced localization method that integrates the Improved Zebra Optimization Algorithm (IZOA). First, a dual communication radius mechanism is introduced to refine hop count estimation and mitigate distance overestimation in dense regions. Then, a bent-path eliminating method with adaptive angle thresholds based on anchor node density is employed to eliminate intermediate anchor nodes on overly curved paths, effectively reducing distortion in average hop distance estimation caused by topological holes. Finally, IZOA is employed to globally optimize the coordinates of unknown nodes. Simulation results demonstrate that, under network topologies containing O-shaped topological holes, the proposed method achieves significantly higher localization accuracy, reducing average localization error by 30%, 10%, 8% and 6% compared to DV-Hop, WBOA-DV-Hop, IWOA-DV-Hop and TGWO-DV-Hop respectively. Showing good accuracy.