For the purpose of source localization, we have proposed two recursive algorithms in our companion paper, which use time difference of arrival (TDOA) measurements received from sensors by accounting for random uncertainties in sensor positions. This paper is devoted to presenting a comparative analysis on the two recursive localization algorithms. The first algorithm is called recursive localization algorithm, which uses the current estimate of source position to form a new measurement equation of the unknown source position. The second algorithm firstly estimates an auxiliary variable and then rearranges the nonlinear TDOA equation into a linear measurement equation. By employing the update covariance of the update localization of the two algorithms, it is shown that the second algorithm outperforms the first one. An illustrative example is included to validate our theoretic results.