Productivity and the need for its development is one of the main challenges that face the construction industry and it is important issue in decision making in the field of construction management. The most reliable measure of productivity is the output per work- time achieved by the labours on construction site. This research aims for developing a model for predicting the productivity of gypsum plastering works for building projects using multiple linear regression techniques. The model was developed based on forty nine set of data collected in Baghdad city from the period (January to June 2012) through observation. In this research, the regression model was developed in four runs by using (MS Excel) program, the result showed that Model four was more fitted model with average accuracy percentage of (95%) to predict productivity of gypsum plastering works in buildings.