Many solutions for battery modeling have been proposed in the literature; they usually rely on the definition of a generic model template (e.g., a circuit equivalent), which is then populated using data obtained from direct measurements on actual devices. This approach incurs in high costs (for the instrumentation) and long characterization times. In this work we present a modeling framework that is an extension of [1], and which allows to automatically build battery models using only data reported in the battery datasheets, without requiring any measurement. Specifically, we extend the existing methodology by including temperature dependence in the battery behavior. The flexibility and the accuracy of the proposed strategy are demonstrated on some sample batteries that provide temperature-dependent information in their datasheets.