A decision procedure to determine when computer software should be released after testing is described. This paper extends optimum release policies minimizing the total expected software cost with a scheduled software delivery time under reliability requirement constraint. Such cost considerations enable us to make a release decision as to when transfer a software system from testing phase to operational phase. The underlying model is software reliability growth model described by a nonhomogeneous poisson process. It is assumed that the penalty cost function due to delay for a scheduled software delivery time is linearly proportional to time. Numerical examples are shown to illustrate the results.