**MSC:**- 62E17 Approximations to distributions (nonasymptotic)

Let $X_{(1)}, X_{(2)},...,X_{(n)}$ be the corresponding order statistics. We assume that for some k and r with $1\leq k\leq r

for some $r\leq s\leq n$ with specified probability. we use a classical approach based on the statistic $T=\displaystyle{\sum_{i=k+1}^{r}X_{(i)}+(n-r)X_{(r)}-(n-k)X_{(k)}}$

we determine a prediction interval for $X_{(s)}$. We have derived maximum likelihood estimates based on doubly type-II censored samples , with these estimated parameter values

we determine a maximum likelihood predictor. We have studied the Bayesian approach when the prior distribution for the parameter is given by a two parameter gamma distribution.