Interval censored data arise when a failure time say, Y cannot be observed directly but can only be determined to lie in an interval obtained from a series of inspection times. Therefore, prediction of unobserved Y can be interesting issue. In this paper, we will obtain the approximate maximum likelihood estimators (AMLEs) of unknown parameters in the Weibull distribution and using them, we introduce a point predictor of Y in the i-th, i=1,2...,n interval, as the expected value of the conditional distribution of Y given L and R. In addition, using the AMLEs of Weibull parameters, the predictor of Y and percentile bootstrap, we will get the 95% confidence bounds of unkown parameters as well as 95% prediction bound of Y in the i-th interval. A Monte Carlo simulation study and an illustrative example presented to assess the proposed performance of the procedures.
