This paper explores the framework and development of stochastic frontier Approach (SFA). The original idea of the SFA and its theoretical framework is discussed to provide a basic foundation of the approach. The development of SFA with more flexible distribution assumptions follows the pioneering model. Experts also
develop the time-variant technical efficiency models, in order to allow variation between times for a production unit. The most recent development is the panel data SFA, which includes the two-stage and the one-stage procedures.