Spatio Temporal or Space Time Model is a stochastic processes which indexed by space and time simultaneously. In this paper we studied a development of spatio temporal model especially in Generalized Spatio temporal Autoregressive (GSTAR) model which is developed from Spatio temporal Autoregressive (STAR) model from Pfeifer (1979). STAR and GSTAR models are developed as a univariate time series from Box-Jenkins (1976). GSTAR model is a stationary multivariate time series model, which has an assumption that parameters are vary per location, so it is capable for heterogenous locations characteristic. GSTAR model can be applied for forecasting an observation at the future time based on lag time before and influenced by the observations at surrounding locations. For example of GSTAR model can be used in forecasting of oil production at oil wells at volcanic field Jatibarang, forcasting of tea productivities at several plantations in West Java and forecasting of rainfall at West Java area etc. The stationary GSTAR model can be extend to be GSTAR-X with addition of exogeneous variable, or to be non-stationary model as GSTARI or GSTAR-ARCH and KSTAR-Kriging. To make easier in estimaton of parameters of GSTAR model , we built an interactive software using the script of opensource R software using Ordinary Least Squares (OLS) method.  Spatio Temporal Model especially the GSTAR model can be used for recommendation of management in decision making at a certain area.