Thin films devices include MEMS gas sensors and thin-films lithium-ion batteries (TFLIBs) are promising technologies for future applications of various electronics. In this talk, we introduce some recent progress in applying numerical approaches to achieve efficient screening of high-performance nano-materials-based thin films devices. The key parameters of gas sensing materials (sensitivity and selectivity) and anodes materials in LIBs (reversible capacity, irreversible capacity, cycling performances) are efficiently predicted and improved. It is demonstrated that the proposed methods show excellent accuracy for predicting functional properties of the devices. The attempts on incorporating machine learning for accelerating the numerical frame are also presented.