The transesterification reaction used to produce biodiesel was monitored with 1H high-resolution nuclear magnetic resonance (HRNMR), conventional low-resolution NMR (LRNMR), and unilateral NMR (UNMR) spectroscopy. HRNMR was used as a standard method to compare with the methods of LRNMR and UNMR. A Carr–Purcell–Meiboom–Gill (CPMG) pulse sequence was used in both LRNMR and UNMR experiments. In LRNMR, the CPMG signal was used to measure the transverse relaxation time (T2), which depended on sample viscosity; it showed a good correlation (R = 0.994) for the concentration of biodiesel in the oil/biodiesel mixture. These measurements could only be used when the samples contained less than 1% of methanol. In UNMR, the CPMG decay of the biodiesel/oil mixture strongly depended on molecular diffusion because of the intrinsic high magnet field gradient (G) associated with the magnetic. The T2eff had a good and negative correlation (R = −0.997) with the biodiesel concentration in the biodiesel/oil mixture. The UNMR measurements were insensitive to the methanol contamination up to 50%, which was contrary to the LRNMR method. As methanol diffused rapidly, its magnetization lost coherency quickly in the presence of strong G and it was not fully refocused by the CPMG sequence. Therefore, the use of the fast, portable, and low cost UNMR sensor to monitor transesterification reactions was demonstrated ex situ. Now, the UNMR sensor has been adapted to monitor the transesterification reaction in situ, in a biodiesel pilot plant.