An adaptive algorithm is presented for extracting the flux of the fiber spectrum from a two-dimensional
image observed by LAMOST (Large Sky Area Multi-Object Fiber Spectroscopic Telescope). The new
algorithm is based on RBF (Radial basis function) neural network, employing the Gaussian basis
function to approximate the profile of the spectrum in the spatial direction. In this study, an experiment
is performed with the simulated data. The experimental results show that the new algorithm can highly
enhance the computing speed while preserving the accuracy in the flux extraction. A feasible approach
is thus offered for extracting the flux of the fiber spectrum for LAMOST.