To address the issue of drastically degraded positioning performance in non-line-of-sight environments due to occluded line-of-sight paths，a three-dimensional positioning and orientation estimation algorithm for millimeter wave single Base Station (BS) system with aid of a single Intelligent Reflecting Surface (IRS) is proposed. Firstly，through leveraging the sparse characteristics of millimeter wave signals and incorporating tensor theory，the problem of estimating channel parameters related to the location information is transformed into a tensor decomposition problem for resolution. Secondly，the estimated positioning-related parameters are used in a hybrid algorithm involving Time of Arrival (TOA)，Angle of Arrival (AOA)，and Angle of Departure (AOD) to estimate the coordinates and orientation of the Mobile Station (MS). To assess the performance of the proposed positioning algorithm，theoretical bounds for the estimation errors of MS location and orientation are derived. Simulation results demonstrate that the proposed algorithm closely approximates the theoretical error bounds，confirming the effectiveness of the proposed algorithm.