Abstract. A new incoherent scatter radar called EISCAT 3D is being constructed in Northern Scandinavia. It will have the capability of producing volumetric images of ionospheric plasma parameters using aperture synthesis radar imaging. This study uses the current design of EISCAT 3D to explore the theoretical radar imaging performance and compares numerical techniques that could be used in practice. Of all imaging algorithms surveyed, the singular value decomposition with regularization gave the best results and was also found to be the most computationally efficient. The estimated imaging performance indicates that the radar will be capable of detecting features down to approximately 90x90 m at a height of 100 km, which corresponds to a ~0.05° angular resolution. The temporal resolution is dependent on the signal-to-noise ratio and range resolution. The signal-to-noise ratio calculations indicate that high resolution imaging of auroral precipitation is feasible. For example, with a range resolution of 1500 m, a time resolution of 10 seconds, and an electron density of 2·1011 m−3, the correlation function estimates for radar scatter from the E-region can be measured with an uncertainty of 5 %. At a time resolution of 10 s and an image resolution of 90x90 m, the relative estimation error standard deviation of the image intensity is 10 %. Dividing the transmitting array into multiple independent transmitters to get at multiple-input-multiple-output (MIMO) interferometer system is also studied and this technique is found to increase imaging performance through improved visibility coverage. However, an estimate shows that this reduces the signal-to-noise ratio. MIMO is therefore only useful for the most brightest targets, such as meteors, polar mesospheric summer and winter echoes, and satellites. The results show that radar imaging of is feasible with the EISCAT 3D radar, and that the use of the MIMO technique should be explored further.