In this study we present a two-stage adaptive estimator of prevalence in the presence of test errors. We assume that tests are not 100% perfect. We obtain the adaptive estimator using Maximum Likelihood Estimate (MLE) method and use Fisher information to determine the variance of the estimator. We use Matlab, a statistical software for simulation and verification of the model. We analyse and discuss the properties of the constructed estimator in comparison with other existing estimators in the literature of pool testing. We also provide the confidence interval of the estimator. When the test kits have low sensitivity and specificity, we establish that the adaptive estimator outperforms other existing estimators. Further more, we demonstrate that the efficiency of the adaptive estimation scheme improves as the number of stages increases. This makes the adaptive testing scheme more ideal in areas where errors are rampant.