Neutron holography is a new imaging technique based on the recording of the interference pattern of two coherent waves emitted by the same source, which allows observing the spatial order of microscopic objects like molecules or atoms in crystal sample. Two approaches can be used in neutron holography measurements. One is called inside-source holography, in which both the reference wave and object wave come from embedded atoms in the sample and propagate toward the detector outside the sample. The second approach called inside-detector holography is the inverse method of inside-source holography, in this case the reference wave is the initial neutron beam coming from a distant source outside the sample, while the atoms embedded in the sample act as detectors. In an ideal inside-source holography experiment, the sample should be fixed and the detector moves on a sphere, which is not practical because the detector system is usually heavy and far from the sample. In order to minimize the operation space, the detector always moves on a circle around sample or is located at a fixed position, while the sample rotates in an appropriate way to imitate the motion of the detector in a sphere. However, the orientation of the sample relative to the incident neutron beam is changed during sample rotation, and part of the inverse hologram is recorded together with the inside-detector hologram, which can cause distortion in the holographic reconstruction. In this paper, we simulate neutron holograms and reconstructions based on three different sample/detector rotations. In the first case, the detector moves on a circle, while the sample rotates about an axis perpendicular to the detector moving surface. In the second case, the detector is fixed, while the sample rotates around two perpendicular axes, the θ axis rotating the sample through πradians is perpendicular to the incident beam-detector plane, while the θ axis rotating the sample through 2πradians moves on a circle parallel to the incident beam-detector plane, this rotation can be carried out on a 3-axis spectrometry. In the third case, the detector is also fixed and the sample rotates around two perpendicular axes, but the θ axis is parallel to the sample-detector direction, while the θ axis moves on a circle perpendicular to the incident beam-detector plane, this rotation can be carried out on a 4-cycle spectrometry. The distortions and corresponding correcting methods of three kinds of rotations are discussed. The result shows that most distortions can be corrected by using special measurement or reconstruction techniques. Furthermore, pure sample rotation based on 3-axis spectrometer can achieve the best reconstruction result, so this rotation approach is preferred if conditions permit.