This paper presents a RANS-based far field broadband noise prediction method for flow generated noise in low Mach number and high Reynolds number, such as slat noise. The compressible Navier-Stokes equations were linearized and reformulated firstly for acoustic generation and propagation. Then the far field noise power spectral density formulation was obtained, which relates the far field noise spectrum to the volume statistics and the exact Green's function that accounts for the sound flow coupling and propagation effects, including the diffraction and scattering effects caused by solid surfaces and the refraction effects by nonuniform flow. The adjoint Green function method was employed for the fast exact Green function computation. A semi-analytical and semi-numerical method was proposed to tackle the nearly singularity problem appeared in the computation of exact Green function. Thus, the homogeneous and isotropic turbulence was assumed for the source modeling based on the mean flow and turbulence statistics from RANS. Finally the method was applied into the prediction of slat broadband noise based on the mean flow of 30P30N multi-element airfoil at four degree of angle of attack under the uniform mean flow assumption. The computation of spatial source distribution demonstrates the downstream moving of the source location to reattachment region and the decreasing of the spatial dimension of source with the increase of the frequency. The comparison between the predicted spectral and the modeled spectral shows a good agreement in peak frequency and the middle frequency range, and demonstrate the accuracy of the method.