The electronic band structure of materials plays a decisive role in their thermoelectric performance. The long-standing focus in this field has been on engineering the band structure to control the power factor and thermoelectric figure of merit of thermoelectric materials. Common methods for band structure engineering include introducing multiple valley degeneracy, creating anisotropic complex Fermi surfaces, and adjusting mobility. These methods have significantly enhanced the thermoelectric performance of materials, underscoring the effectiveness of band structure engineering in the design and optimization of thermoelectric materials. However, for materials with specific structures or compositions, determining and achieving the corresponding optimal band structure for optimized thermoelectric performance is not a straightforward task and often requires extensive trial-and-error experiments or calculations. In theory, first-principles calculations can provide insight into the relationship between band structures and atomic orbitals and their interactions, thereby elucidating the ways to engineer band structures. However, due to the multitude of parameters involved and their interdependencies, it is challenging to directly establish the relationship between specific orbital effects and thermoelectric performance through calculations alone. Recently, a team from the National Institute for Materials Science and Tsukuba University in Japan proposed a method called orbital sensitivity analysis to elucidate the relationship between orbital interactions and band structures, thereby achieving thermoelectric performance optimization. This proposal is primarily based on the tight-binding model of band structures. For a specific material, its band structure can be described by a series of highly complex orbitals and their interaction parameters. Typically, these parameters are numerous and correlated with each other. The researchers first developed a symmetry-adapted independent parameter search model, which can automatically identify a few parameters that can be independently controlled for a specific material. They then constructed the relationship between these parameters and critical band features and thermoelectric transport properties through sensitivity analysis. Taking the typical thermoelectric system PbTe as an example, the researchers validated the effectiveness of this method. The results showed that the transport properties of PbTe are primarily determined by only a few orbital parameters. By manipulating certain orbital parameters, such as the orbital interaction between Te px and Te py, it is possible to achieve the approximate degeneracy in the valence band orbitals, thereby tripling the system's power factor at high temperatures. At low temperatures, adjusting the parameters to favor the light band at the band edge is beneficial for enhancing the power factor. These theoretical findings provide directions for experimental control, such as stress application direction and doping manners. The method proposed in this study can comprehensively establish the relationship between material orbital interactions and transport properties, and can also be easily extended to other systems. Therefore, it holds the potential to become a powerful tool for optimizing the performance of thermoelectric materials.