Precise Time Of Arrival (TOA) estimation is a basic step of standard migration methods for object imaging from SAR measurements. In this paper, an effective computation method of the TOA for through wall model of object recognition is presented. The conventional method that uses constant velocity model produces errors in object shape and position estimations. Computation of the TOA (corresponding to true flight distance) for three layer model requires the complex minimization algorithm. Proposed method transforms three layer (air-wall-air) model to equivalent two layer (air-wall) model with lower computation complexity. It uses iterative solution of well defined minimization problem. Moreover, conveniently selected initial conditions of iteration process can further decrease computational complexity of the method. The proposed method provides more precise TOA estimation than conventional one and is less complex than three layer methods. Therefore, it is suitable for implementation on realtime hardware. The method performance is demonstrated by processing of real 2-dimensional SAR data acquired by through wall M-sequence UWB radar system.