Sound analysis has become an important tool for studying the behavior, health and welfare of animals in recent years. Monitoring laying hens’ sound can be used to evaluate their health or welfare. However, large-scale henhouses exist more noise interference such as a typical fan noise, which has a great effect on accurately recognizing poultry vocalization. Hy-line brown laying hens were taken as an example and poultry sound signal pretreatment was applied for reducing the fan noise interference as far as possible. Different types of laying hens’ vocalization and fan noise were recorded by digital sound collection platform. LabVIEW software was used for sound signal processing and analyzing time-frequency characteristics of the fan noise as well as poultry vocalization. Meanwhile, compared with different noise reduction methods, including IIR filter de-noising, wavelet threshold de-noising and improved spectral subtraction methods, and the effect of different de-noising methods were evaluated. Results showed that the sound frequency of laying hens were mainly distributed between 400Hz and 2500Hz, and fan noise frequency distribution was between 0Hz and 1000Hz. In the acoustic environment of SNR within -8~20dB, the improved spectrum subtraction method had the minimum root mean square error (0.03~0.38) with the shortest processing time (6~7ms), which had good outcome in practical application. This study provided foundation for analyzing laying hens’ vocalization in fan noise environment of large-scale henhouses.