The research team has deployed a monitoring system for assessing the nuisance levels in veterinary clinics by unsupervised learning and deep learning of Long-Sort term memory (LSTM) model. The monitoring unit used MQ-series gas and DHT11 sensor for measuring odor intensity, ambient temperature, and humidity. The collected data were stored in Cloud, preprocessed, classified, and subsequently used as training data for the deep learning LSTM model to predict nuisance levels categorized into four classes: L0-low level, L1-medium level, L2-high level, and L3-very high level. The analysis results indicate that LSTM model demonstrates effective performance in odor pollution prediction, reflected by evaluation metrics such as recall, f1-score, and accuracy, as well as error metrics including RMSE and explained variance score (EVS).