Cracked concrete structures can self-heal within certain limits of crack width and water pressure, primarily through the formation of CaCO3. However, construction practice has shown that self-healing is not always as effective as expected by codes, indicating that the boundary conditions for effective self-healing are only partially understood. In this study, cracked concrete specimens made with CEM I or CEM III/A are subjected to a self-healing test according to Edvardsen after 7 or 28 days of curing to validate previous research results and to investigate the healing performance as a function of Portland cement content and degree of hydration. The measurement of the volumetric flow rate over time serves as an indirect measure of the healing efficiency. Generally, the flow rate decreased exponentially. For CEM I samples, the healing efficiency was independent of the degree of hydration. For CEM III/A samples, a similar self-healing performance within the first days of the healing process was observed. At later stages, however, the healing efficiency was significantly lower compared to CEM I. CEM III/A samples cured for 7 days, even exhibited a slow increase in volumetric flow rate after about 30–40 days of healing.