In the present study, a computational analysis of the flow in a centrifugal blower is carried out to predict a performance and to explain noise characteristics of the blower. Unsteady, 3D Navier-Stokes equations were solved with k-${varepsilon}$ turbulence model using CFX software. CFD results were compared with the experimental data that is acquired from an experiment conducted with the same blower. The pressure fluctuation in the blower was transformed into the frequency domain by Fourier decomposition to find the relationship between flow behaviors and noise characteristics. Sound pressure level (SPL) which is obtained from wall pressure fluctuation at impeller outlet represents relative overall sound level of the blower well. Sound spectra show that there are some specific peak frequencies at each mass flow rate and it can be explained by flow pattern.