Fatigue tests reaching design-relevant numbers of cycles are highly time-consuming and costly, resulting in a limited availability of experimental data. Therefore, reliable methods are required to predict the number of cycles to failure from deformation-based fatigue parameters during an ongoing fatigue test, without requiring specimen failure. This would enable fatigue tests to be terminated early while still providing reliable fatigue life predictions, thereby reducing experimental effort and improving the utilization of available experimental data. This study investigates the suitability of two deformation-based fatigue parameters, namely the strain rate at the maximum compressive stress level and the cumulative dissipated energy, for fatigue life prediction. To this end, 30 single-stage cyclic compression fatigue tests on high-strength concrete were systematically evaluated at two compressive stress levels. Fatigue life was predicted using data obtained at different stages of the fatigue tests, enabling an assessment of both the predictive capability of the investigated parameters and the minimum test progress required for a reliable prediction. Significant correlations were found between the number of cycles to failure and both fatigue parameters (R2 > 0.99). The strain-rate-based approach consistently outperformed the dissipated-energy-based approach and provided the most accurate fatigue life predictions. Furthermore, reliable predictions of the number of cycles to failure were achieved with 50% of the fatigue life, demonstrating that deformation-based fatigue parameters enable the efficient utilization of incomplete fatigue experiments and have the potential to substantially reduce experimental effort in future fatigue investigations.