Fully developed turbulent pipe flow is investigated using two methods, magnetic resonance velocimetry (MRV) and Large Eddy Simulation (LES). Magnetic resonance velocimetry is a non-intrusive flow visualization technique, which is able to measure the 3 dimensional 3 component (3D3C) mean velocity field of complex flows using an MRI scanner. In the case of MRV, the partial volume effect is one of the main sources of error in the measured velocity at the solid wall boundary. This causes uncertainty in the integrated flow rate. In the present study, we evaluated the measurement uncertainty of MRV resulting from this partial volume effect in a fully developed turbulent pipe flow with Reynolds number of 11,700. We introduced a correction procedure to reduce this error contribution. In addition, the same pipe flow is simulated by LES using an immersed boundary method. Both the experimental and computational approaches are validated by comparison with a direct numerical simulation (DNS) data set, showing that the velocity profiles obtained by both approaches agree well with the reference data set.