Sensor-less vector control of Surface Mount Permanent Magnet Synchronous Motor (SPMSM) throughout the entire speed regime is a challenging problem in PMSM drive. This paper addresses this control problem and presents the design and simulation study of sensor-less vector control of SPMSM using Cubature Kalman filter (CKF) based rotor position and speed estimator. The use of CKF for speed and position estimation of PMSM is reported in recently literature. However, the results are not substantial. Position and speed estimation errors are quite large, also the estimated speed and position contain considerably large ripple. In the present work the performance of CKF observer in terms of position and speed estimation accuracy is significantly improved. The dynamic performance of the drive throughout the entire speed zone under variable load torque is also improved.