Ab initio molecular electronic structure theory has been employed in order to investigate systematically the X 3B1, a 1A1, b 1B1, and c 1Sigmag+ states of NH2+, with emphasis placed on the b 1B1 and c 1Sigmag+ states. The self-consistent-field (SCF), configuration interaction with single and double excitations (CISD), complete active space (CAS) SCF, and CASSCF second-order configuration interaction (SOCI) wave functions with nine basis sets, the largest being a triple-zeta basis set with three sets of polarization functions and two additional sets of higher angular momentum and diffuse functions [TZ3P(2f,2d)+2diff], were used to determine equilibrium geometries, harmonic vibrational frequencies, infrared (IR) intensities, and dipole moments. The ground, first, and second excited states are confirmed to be bent, while the third excited state is predicted to be linear. The bond angles of NH2+ are shown to be larger than those of the corresponding isoelectronic CH2 molecule. At the highest level of theory, TZ3P(2f,2d)+2diff CASSCF-SOCI, the triplet-singlet splitting is predicted to be 29.4 kcal/mol (1.28 eV, 10 300 cm-1), which is in good agreement with the experimental observation of 30.1 kcal/mol (1.305 eV, 10 530 cm-1). With the same method, the second excited state (b 1B1) lies 43.7 kcal/mol (1.89 eV, 15 300 cm-1) above the ground state, which is significantly lower than the experimentally proposed value of 2.54 eV. The third excited state (c 1Sigmag+) is predicted to lie 77.0 kcal/mol (3.34 eV, 26 900 cm-1) above the ground state. The equilibrium geometry of this c 1Sigmag+ state is determined to be re = 1.030 Angstrom at the TCSCF-CISD level with the largest basis set. Since the IR intensities of all active vibrational modes are predicted to be substantial, IR spectroscopic studies of the four states are feasible. However, of the six fundamentals experimentally assigned to date, two appear to be incorrect. The energy separations among the four lowest-lying states of NH2+ are found to be larger than the corresponding states of CH2.