Objective: To enhance the accuracy and efficiency of tumor ablation procedures, this study integrated tumor ablation technology with a surgical navigation system and addressed errors induced by electromagnetic interference in clinical settings. Methods: An improved singular value decomposition (SVD) algorithm was proposed, which iteratively traversed the point set by selecting three corresponding points at a time to compute transformation matrices. Each matrix was applied to the original dataset, and the one yielding the minimum error between the transformed and target point sets was selected as the optimal solution. This approach mitigated the influence of individual outlier points and significantly reduced navigation errors caused by electromagnetic interference, thereby enhancing registration accuracy and robustness. Results: Experimental results demonstrated that the improved algorithm achieved high accuracy in simulating lesion localization. The root mean square error (RMSE) was employed as a quantitative metric. RMSE values were 0.85, 0.82, and 0.75 at rotation angles of 0°, 5°, and 10°, respectively, indicating minimal deviation between the transformed and target point clouds. Conclusion: Compared with the conventional SVD algorithm, the improved method yielded consistently lower RMSE values, confirming its effectiveness in minimizing electromagnetic navigation errors and enhancing registration accuracy.