Surface-dependent field-of-view (FOV) inhomogeneity in confocal laser scanning microscopy (CLSM) has been demonstrated but remains insufficiently quantified, particularly for surfaces with complex, multi-scale geometries. This limitation affects the traceability and comparability of optical surface topography measurements. In this work, we show that FOV-induced measurement deviations are fundamentally surface-dependent and cannot be explained by local slope alone. Using deterministic and stochastic surfaces with controlled slope distributions and spatial frequency content, we demonstrate that smooth geometries exhibit negligible FOV effects, whereas complex high-frequency surfaces exhibit pronounced position-dependent distortions. To interpret this phenomenon, a surface-complexity-driven framework is introduced, linking the angular transfer characteristics of the instrument to the surface slope distribution via a complexity parameter î. By relating î to gradient magnitude distribution descriptors (e.g., Sdq95) and spatial frequency characteristics (e.g., Sal), the framework provides a physically grounded interpretation of FOV-dependent measurement limits. The results indicate that, for complex surfaces, CLSM measurements become increasingly instrument-limited rather than surface-limited, due to position-dependent loss of high-angle information. The observed FOV effects are consistent with scanning-induced angular limitations, including pupil shift and/or clipping, and off-axis aberrations. This work establishes a foundation for task-specific, traceable optical metrology of complex surfaces.