Due to its ability of optical sectioning and low phototoxicity, z-stacking light-sheet microscopy has been the tool of choice for in vivo imaging of the zebrafish brain. To image the zebrafish brain with a large field of view, the thickness of the Gaussian beam inevitably becomes several times greater than the system depth of field (DOF), where the fluorescence distributions outside the DOF will also be collected, blurring the image. In this paper, we propose a 3D deblurring method, aiming to redistribute the measured intensity of each pixel in a light-sheet image to in situ voxels by 3D deconvolution. By introducing a Hessian regularization term to maintain the continuity of the neuron distribution and using a modified stripe-removal algorithm, the reconstructed z-stack images exhibit high contrast and a high signal-to-noise ratio. These performance characteristics can facilitate subsequent processing, such as 3D neuron registration, segmentation, and recognition.