Abstract: Ring-coded aperture imaging system is used to diagnose low intensity pulsed radiation source, which performs high resolution and high signal-to-noise ratio (SNR). Through Geant4 Monte Carlo toolkit, this paper simulates the system to get six point spread function (PSF) of point sources centered at different distance and the coded image of neutron source made up of a group of lines. Sectioned restoration algorithm is one of the methods to solve image restoration of space-variant point spread function. Firstly, the reconstructed image is divided to rectangle section or circle section. Each sub-section is restored by RL iterative method. Then, the boundary regions which bring obvious noise are removed. Those pixels are superposed by adjacent sections with weighting coefficient, which is the proportion of the distance between pixels and boundary. At last, the sub-frames are spliced together to construct integrated image. Computer simulations indicates the modified method is effective for removing the boundary distortion, which can diagnose the distribution of radiation source.