<p>The objective of this study was to investigate the impact of structural elements in 3D respiratory sensors on respiratory signal detection performance, focusing on variations in the sensor's contact surface area. To achieve this, a range of experimental variables was selected, including the size of 3D-printed fillers (small, medium, and large), the elasticity of the base fabric (woven versus elastic bands), pattern shapes (zigzag, moss, and pile), and pattern layout orientations (vertical and horizontal). By systematically combining these variables, a total of 36 wearable chest belt-type 3D-structured respiratory sensors were fabricated. Respiratory signal detection experiments were conducted using these chest belt sensors on a dummy model and an adult male subject, adhering to a defined experimental protocol. The performance of the 3D-structured sensors was evaluated against key metrics, such as signal homogeneity, repeatability, and reproducibility, with an emphasis on the morphological characteristics of the signal waveforms. Furthermore, the accuracy of the detected signals was quantitatively assessed using the baseline drift index, double-peak index, and correlation analysis with the reference signal SS5LB (BIOPAC Systems). The results of both the pilot and main experiments demonstrated that most 3D-structured sensors effectively detected respiratory signals. Among the tested configurations, the most homogeneous, repeatable, and reproducible respiratory signals were observed when moss or zigzag patterns in the 3D structure were configured in a horizontal layout. Specifically, the moss pattern in a horizontal layout achieved the highest correlation coefficient with the SS5LB reference signal (r² = 0.80), validating this design configuration as the most effective combination of variables for respiratory signal detection.</p>