Ab initio molecular orbital theory has been used to investigate the barrier to inversion in NF3+. Complete geometry optimizations have been carried out with a variety of basis sets and electron correlation methods. It is shown that for this problem the use of single-point energy calculations needs to be treated with caution. Discrepancies amongst previous theoretical values are shown to be due to the use of inappropriate geometries and inadequate inclusion of electron correlation. With a variety of quantum chemical methods, the barrier to inversion is predicted to be in the range 11.3-13.3 kcal/mol. We have also predicted the vibrational splittings for the NF3+ potential energy surface and attempted to reconcile the differences with experiment.