Abstract: This study investigates the adaptability of the Gaussian filtering method in non-contact measurement of surface roughness. Using surface profile data obtained from a spectral confocal displacement sensor as an example, the Gaussian filtering algorithm is introduced to compute the filtering centerline and separate the roughness signal. A program for the Gaussian filtering algorithm is designed on the LabVIEW platform, calculating the roughness values and comparing them with the results of current contact calibration. Experimental results indicate that the Maximum Permissible Error (MPE) of the R_a value after non-contact data processing is ±5%, demonstrating that the Gaussian filtering method is suitable for data processing in non-contact measurements. The modular design of the algorithm in LabVIEW is easy to implement, integrates well with other modules, is easy to maintain, and holds significant potential for wider application.