Reliability of Li-ion batteries (LIB) is major concern for FHE devices due to needs of flexibility without degradation of state of health (SOH) of the LIB. In this regard, a thin form factor based LIB below 1mm of thickness is regarded as the candidate material to meet such needs because it is able to be fold, bent and twisted with limited performance drop. In addition to this, LIB has high specific power (W/Kg) and high specific energy (Wh/Kg) and a lower memory effect which could make the LIB more attractive for wearable applications. While studies on chemo-physical effect such as SEI growth, material decay, etc. due to repeated charging and discharging LIB has been conducted greatly, but such effects due to the flexing LIB have been rarely conducted. In this study, degradation of thin-flexible power source reliability has been studied under twist, flexing, flex-to-install of magnitude to replicate stresses of daily motion of human body using motion-control setups in a lab-environment. Additionally, AI-based regression model has been developed to predict the SOH of the battery with multiple variables including physical, ambient and chemo-mechanical experimental conditions which could be challenged to be treated by the manpower. The developed models can be used to predict the life of the battery and analyze acceleration factors between test conditions and use conditions for variety of test conditions based on the individual variables and their interactions.