This research delves into the software testing process and its environmental factors to uncover the core elements influencing software reliability. Specifically, it focuses on the learning and negligent aspects of the software reliability growth model. The learning factor accelerates reliability growth, leading to an S-shaped curve in the mean value function, while the negligent factor highlights the occurrence of imperfect debugging. The study also uses Brownian motion and stochastic differential equations to establish statistical confidence intervals for reliability and costs. These intervals aid software managers in assessing potential release risks at various confidence levels, allowing them to make informed decisions considering resource constraints and desired system reliability levels across different scenarios.