Software process modeling and simulation hold out the promise of improving project planning and control. However, purely quantitative approaches require a very detailed understanding of the software project and process, including reliable and precise project data. Contemporary project management defines the success of project as a box or hyper-cube, rather than the traditional single point, which allows the planning for software project success semi-quantitatively with uncertainty-tolerance. This paper introduces semi-quantitative reasoning into software project planning and develops a practical approach to enhance the confidence of project success under the uncertainty and contingency. We illustrate its value and flexibility by a simplified software process model focusing on staffing issues.