This paper applies model predictive control (MPC) to multi-input spacecraft attitude maneuver problems using reaction control systems (RCS) and reaction wheels (RW). For this purpose, we consider linear quadratic cost functions under the input constraints of nonlinear on-off control inputs for RCS and linear continuous inputs with saturation for RW. These are optimized as mixed integer quadratic programming (MIQP) problems. We further analyze the closed-loop stability in the input-to-state stability framework and derive a stability condition described by a linear matrix inequality (LMI). Some numerical studies are performed and the effectiveness of the proposed model predictive control is shown.