The increasing demand for petroleum products and the limited supply of fuels in certain parts of the world today have led researchers to focus on alternative fuels. Among the non-edible plant sources for biodiesel production are microalgae, with an oil content of 25-75%. Most oil compounds include palmitic, stearic, oleic and linoleic fatty acids, which can be converted to biodiesel. Soxhlet is the most common method for extracting oil, but it is not suitable because of the toxicity of the solvent, the high solvent consumption, and the complete failure of the oil. The application of supercritical fluid in the extraction of a plant extract is a new method, which is nowadays considered in isolation processes due to its environmental, scientific and economic benefits. Therefore, supercritical carbon dioxide was used to extract oil from native Iranian paracellella microalgae. In the present work, experiments were designed to perform the separation process under different operating conditions using a centralized response procedure and composite design. For this purpose, the effect of four variables of temperature, pressure, supercritical fluid flow intensity and dynamic time in the range of 35-75 ° C, 110-270 bar, 0.8-2 ml / min and 50-130 min respectively on efficiency and recovery rate Extraction is investigated. Transesterification was used to produce biodiesel from the extracted oil, whereby the fatty acids were converted to methyl esters by the catalyst. In this study, we used a catalyst type of catalyst, which showed the use of alkaline catalyst in the transesterification reaction. The results of the extraction process and the transesterification method are presented in terms of extraction efficiency, oil recovery percentage and biodiesel (total fatty acid methyl ester). Experimental results in oil extraction showed a good fit with a detection 93.01% and a modified detection 89.86%. The methods used for evaporating oils at 56.82 ° C, 18.27 MPa, 0.96 ml / min flow rate and 83.94 min dynamic time achieved optimum efficiency. Calculated 44.89%. In this study, artificial neural network was used to model the extraction process. In this modeling performed with forward perceptron neural network with LM training algorithm with 6 hiden layer neurons and validation ratio and 30% input data testing, total detection coefficient was 0.93 which was in good agreement with data. It was experimental. The yield of biodiesel after reaction was approximately 98%. According to this study, microalgae obtained from the Persian Gulf is a very good source of biodiesel production in Iran. Keyword : Paracellella microalgae, biodiesel, supercritical fluid, response surface methodology