Detecting and modelling Land Cover are crucial for managing natural resources, assessing environmental impacts, and preserving ecological connectivity. Convolutional Neural Networks (CNNs) have become pivotal in advancing remote sensing applications, particularly in land cover classification. Therefore, this study introduces a CNN classifier for the prediction land cover efficiently. The preprocessing is carried out based on Principle Component Analysis (PCA), which reduces the dimensionality of the input data while retaining essential information. This step aims to enhance computational efficiency and improve the CNN's ability to extract relevant features during training. Following preprocessing, a comprehensive training phase involves optimizing the CNN architecture using a labelled dataset. This process allows the network to learn intricate spatial patterns and spectral characteristics crucial for accurate land cover classification. Evaluation of the trained model occurs in the testing phase, where performance metrics such as accuracy, precision, recall, and F1-score are computed using an independent validation dataset. Results obtained from the python software demonstrate that the CNN-PCA framework's capability to achieve high classification accuracy across diverse land cover types.