The feasibility of using MR thermometry for temperature feedback to control a transurethral ultrasound heating applicator with planar transducers was investigated. The sensitivity of a temperature-based feedback algorithm to spatial (control point area, slice thickness, angular alignment) and non-spatial (imaging time, temperature uncertainty) parameters was evaluated through numerical simulations. The angular alignment of the control point with the ultrasound beam was an important parameter affecting the average spatial error in heat delivery. The other spatial parameters were less influential, thus providing an opportunity to reduce spatial resolution for increased SNR in the MR imaging. The update time was the most important non-spatial parameter determining the performance of the control algorithm. Combined non-spatial and spatial parameters achieved acceptable performance with a voxel size of 3 mm x 3 mm, a 10 mm slice thickness and a 5 s update time. Temperature uncertainty of up to 2 degrees C had little effect on the performance of the control algorithm but did reduce the average error slightly due to a systematic, noise-induced overestimation of the boundary temperature. These simulations imply that MR thermometry performed on clinical 1.5 T imaging systems is of sufficient quality for use as thermal feedback for conformal prostate thermal therapy with transurethral ultrasound heating applicators incorporating planar transducers.