Detection of lung disease is an emerging field, which is mainly done through an auscultation sound (breathing, wheezing, etc.). This research aimed to develop a deep learning model composed of a CNN and an LSTM network used to classify lung diseases from Lung Auscultation Sounds. These, transmitted into the MFCC, were taken into the model for feature extraction and temporal analysis. Architectures used here include Convolutional Neural Network (CNN) layers applied to detect local patterns in the input data, and Long Short-Term Memory (LSTM) layers to extract the temporal dependencies within sound samples. This hybrid architecture enables the model to accurately classify such lung conditions as Pneumonia, Bronchiolitis, URTI, COPD, and Healthy. The model went through public datasets training and validation resulting in promising results concerning the classification accuracy and generalization. Future research should concentrate on this model being integrated into a hardware-focused system, for example, a smart stethoscope, for the online determination of lung diseases. The proposed approach aims to design a more accessible and efficient tool for early identification and surveillance of respiratory diseases.