This work proposes a novel approach for state of health estimation of lithium-ion cells by developing a capacity fade model with temperature and Ah throughput dependencies. Two accelerated life cycle testing datasets are used for model calibration: a multi discharge rate dataset of an NMC/graphite cylindrical cell and a multi temperature dataset for an LCO/graphite pouch cell. The multi discharge rate dataset has been recorded at 23 °C and for 4 discharge-rates (C/4, C/2, 1C and 3C). The multi-temperature dataset considers the accelerated ageing of the cells at 4 temperatures (10, 25, 45 and 60 °C). An Arrhenius model is chosen for describing the temperature dependency while a power law model is chosen for cycle (Ah throughput) dependency. The model shows a good agreement with experimental data in each analyzed condition, allowing a precise description of the capacity degradation over time. From the single-temperature analysis, it is found that the activation energy decreases with respect to the C-rate: this is due to the fact that at higher C-rates, the irreversible chemical phenomena accelerate, leading to an overall faster ageing of the cell. From the multi-temperature analysis, the power law coefficient shows a quadratic dependency relative to temperature: a minimum for the power law coefficient is found corresponding to 25 °C, due to the fact that both for lower and higher temperatures, the ageing mechanisms are accelerated. Finally, an analysis of the impact of fast charging on cell ageing, in different charging scenarios is provided: the fast degradation of the cells at very low temperatures highlights the importance of an appropriate cooling of the battery during charging operations. This empirical methodology can be easily integrated in battery management system algorithms due to the easiness of the calibration and the low calculation time.