Lithium-ion batteries have become indispensable parts of our lives for their high-energy density and long lifespan. However, failure due to from abusive usage conditions, flawed manufacturing processes, and aging and adversely affect battery performance and even endanger people and property. Therefore, battery cells that are failing or reaching their end-of-life need to be replaced. Traditionally, battery lifetime prediction is achieved by analyzing data from current, voltage and impedance sensors. However, such a prognostic system is expensive to implement and requires direct contact. In this study, low-cost thermal infrared sensors were used to acquire thermographic images throughout the entire lifetime of small scale lithium-ion polymer batteries (410 cycles). The infrared system (non-destructive) took temperature readings from multiple batteries during charging and discharging cycles of 1C. Thermal characteristics of the batteries were derived from the thermographic images. A time-dependent and spatially resolved temperature mapping was obtained and quantitatively analyzed. The developed model can predict cycle number using the first 10 minutes of surface temperature data acquired through infrared imaging at the beginning of the cycle, with an average error rate of less than 10%. This approach can be used to correlate thermal characteristics of the batteries with life cycles, and to propose cost-effective thermal infrared imaging applications in battery prognostic systems.