Computation of desired thermal dose can improve the safety and effectiveness of thermal therapeutic interventions while fast bio-heat in silico simulation can enhance the real-time surgical feedback. However, the existing methods are either computationally expensive or numerically inaccurate. This article presents a new methodology for fast and accurate desired thermal dose computation. New iterative algorithms are developed to adjust control parameters to efficiently obtain the desired lesion volumes, based on efficient solutions of bio-heat transfer in soft tissue. The variation of thermal dose is investigated using single- and multi-lesion configurations, and convergence analysis is presented. The clinically relevant scenario is demonstrated using a patient-specific simulation of focused ultrasound treatment of hepatic cancers to achieve multi-lesion planning (consistent and common lesion volumes in our case) for predictable and repeatable heat exposure sequence. The proposed method provides a common approach for fast iterative computation of single- or multiple-desired thermal lesions.