Coconut is one of the common plants in Southeast Asia countries, which could further develop into various products. This study is aimed to study the effect of soaking conditions (types and concentrations of sugar) on the hardness and acceptance score of freeze-dried meat of “Sáp” coconut variety in Vietnam. Artificial neural network (ANN) model was used for predicting and modelling. A total of 540 datasets with two input variables and two output responses were divided into 70% of the training dataset, 15% of the validation dataset, and 15% of the testing dataset. ANN model was successfully forecasted the impact of soaking conditions on physical hardness and eating quality of product. The high agreement between actual and ANN’s predicted values was found with high coefficient determination (R2 > 0.9). Moreover, the correlation between the hardness and acceptability of product also provided. This information could be supported for further study to optimize the process of producing high-quality and sensory freeze-dried coconut meat.