Finite field multipliers found extensive use in the part of cryptography for the purpose of scalar multiplication. The output of those multipliers may have faults, which will lead the cryptosystem to malfunction. This paper introduces a Fault Tolerant technique, which should be capable to pick up the partial failure without affecting the overall performance. The general approach to fault tolerance is redundancy technique. The technique used here is based on Concurrent Error Detection (CED) namely time redundancy. In time redundancy, Recomputing of Duplication with Comparison method is used to detect the errors. This method has been implemented in a standard Montgomery Multiplier. The outputs of the original input and duplicated inputs are compared and checked for the error detection. If error exists, it is corrected using a simple bit invert. The simulation results are obtained using the Modelsim 1O.Ob and its area and power are obtained using Xilinx ISE 9.li.