This paper is about developing a microcontroller-based system that will extract the sound features of post harvested coconuts (cocos nucifera) and
with machine learning to determine the level of maturity. The microcontroller-based system is used as the data gathering feature with sound
acquisition application, conversion, and digital signal processing for machine learning operation. The sampling rate used is 44.1 kHz. The three
maturity stages for the coconut maturity considered are the immature, mature, and over-mature. The machine learning algorithm considered are
Artificial Neural Network, Support Vector Machine, and Random Forest Classifier to perform classification and prediction of the coconut samples
according to its maturity level. The data partition of 70%, and 30%of the whole samples are used, as the training dataset and the testing dataset,
respectively. The data feature in time-domain and frequency-domain shows that the time-domain shows better results than the frequency-domain.
The data partition 70-30 percentage performed well, where the three machine learning algorithms have an above 80% accuracy. The ANN accuracy
for training and testing was 82.3% and 85.21 %, the SVM accuracy was 92.1% and 80.86%, and the RF Classifier was 92.48% and 84.34% with the
data used in time-domain.