Background: Artificial neural network (ANN) is an optimization method which is able to interact the input data and predict the best outputs. The ANN model determines the important factors affection the process. This could allow maximization of the process outputs. Objective: There was only very limited publication on use of ANN for optimization of the bioprocess. This was a trial to use the model to optimize an extraction process and relate the extraction to the antimicrobial activities. Methods: An artificial neural network as model was tested to optimize the extraction for clove flowers and relate it to the antimicrobial activities of the extracts. ANN model was constructed as by multilayer perception (MPL) with six input, two hidden layer and one output layer. The mean for the inhibition zone was 1.5 cm so the data categorized into two sets. Large inhibition zone > 1.5 – 2.3 cm and intermediate or small inhibition zone <0.7-1.5 cm. The antimicrobial activities were tested against 20 microbial strains including 16 bacteria and 4 fungi. Six extraction method were performed and the resulted extracts were examined for their antimicrobial activities against 20 microbial strains which resulted in 120 readings. The statistical calculations for the model outputs were performed. The inhibition zones were taken as preliminary results for the model however the MIC was determined. Results: The model shows an excellent performance with overall accurate prediction 86.6 %. According to the outputs from the model, the most important factor was the use of hot water as solvent for extraction. Calculation of the importance found the hot water important by about 100%. All statistical parameters like AUC, accumulation gain curve and lift chart for the model suggested that the model has an acceptable prediction efficiency. The AUC value was 0.91. GC-Mass was used to identify the most active constituent, it most cases 24 eugenol was identified as active ingredients and some other phenolic compounds Conclusion: According to the results and the statistical calculation for accuracy of the model it was apparent that the model can used successfully to predict and optimized the data for the extraction process. This could incite the ability of these models to be applied on similar processes. The model was efficient in prediction and determination the effective factors.