In the present study, approximate Bayesian confidence intervals for the variance of a normal population under two different loss functions have been derived. The loss functions that are employed are the square error and the Higgins-Tsokos loss functions. Based on the above numerical results we can conclude the following: