The software measurement is an integral part of the software engineering process. With the rise of the software system and their complexity distributed through diverse development phases, the software measurement process has to deal with more management and performance constraints. In fact, the current software measurement process is fixed and manually planned and has to manage a huge amount of data resulting from the complexity of the software. Thereby, measuring software becomes costly and heavy. In addition, the implementation of the measures is dependent on the developer and reduce the scalability, maintainability and the interoperability of the measurement process. It becomes expert dependent and thus more costly. In order to tackle these difficulties, first, we propose in this paper a formal software measurement implementation model based on the standard measurement specification SMM. Then, a software measurement plan suggestion framework based on a learning-based automated analysis