Based on the experimental study,it was found that melamine solution excited by UV light can generate a strong fluorescence. The fluorescence spectrum is within a range from 310 to 600 nm,the peak wavelength of the fluorescence is about 420 nm,and the relationship between fluorescence intensity and melamine solution concentration is nonlinear. A method for the determination of melamine solution concentration was presented,which was based on fluorescence spectroscopy and radial basis function neural networks. For each sample,30 emission wavelength values were selected,the fluorescence intensity corresponding to the selected wavelength was used as the network data,and a radial basis function neural network was trained and constructed. The trained radial basis function neural network was employed to predict the melamine solution concentration in five kinds of samples,and the relative errors of the results were 0.93%,0.09%,0.31%,1.55% and 4.61%,respectively. The results show that this method can determine the content of melamine quickly and accurately. The whole research outcomes will provide a new method for determining the content of melamine and food safety supervision.