Thailand is an agricultural country in Southeast Asia. She always faces the same problems as other developing countries, such as high growth rate of population, low income per capita, unequal income distribution, and low productivity of labor. Thus, government attempts to solve these problems through the national development planning. The first national development planning was adopted in 1960. It did not achieve as expected due to lack of good statistical data and trained manpower. The Ministry of Agriculture and Cooperatives was assigned to make a plan for agriculture. In 1973, there was a joint project between the Ministry of Agriculture and Cooperatives (through its division of agricultural economics), Iowa State University, and United States Overseas Mission/Thailand. The objective of this project was to apply agricultural economic research in supporting development of Thailand's Fourth Five-Year Development Plan for the period 1977-1981. National linear programming and other analyses were constructed except a macroeconometric model. Until 1974, it was felt that the macroeconometric should be included in order to obtain a complete picture of Thai economy. Therefore, the Thai macroeconometric model was constructed. When the model was built, there were some difficulties in terms of econometrics and computer programming. Thus, the equations in the model were estimated by Ordinary Least Squares. Two Stage Least Squares could not be applied directly. These problems led to further study in this dissertation. The objective of this research is to construct a model which can describe the Thai economy in a more extensive and disaggregate manner.;The estimated model in this study consists of 91 equations which contain 68 behavioral equations and 23 identities. The model contains 89 endogenous variables and 107 predetermined variables. The estimation of the model was done by the method of two stage principal component (2SPC). Short time series data from 1963 to 1978 were used in the estimation of coefficients in all sectors except in the compensation of employee equation where the data used range from 1967 to 1978. The model simulation uses the Gauss-Seidel algorithm procedure.;In the model, there remain weaknesses, especially in the investment sector, income distribution sector, and the balance of payment equation. Therefore, the model should be revised. It is hoped that these weaknesses can be reduced in the future.