Providing a robust and reliable model is essential for the traditional way of farming, including crop selection with soil characteristics. The current research develops an emerging evolutionary data-intelligence model: back-propagating neural network (BPNN) integrated with principal component analysis (PCA) to show the performance of the crop classification in Ethiopia. k nearest neighbor (KNN), random forest(RF), decision tree(DT), and support vector machine(SVM) models were also employed to compare the classification performance. For this purpose, take input data with the corresponding cabbage, cowpea, soybean, sesame, groundnut, tomato, onion, and maize as the target variables obtained from the crop were used. The classification models are evaluated based on the three numerical indices, namely error value, accuracy, and performance metrics. To examine the similarities and differences between the observed and predicted values, generate the two values using a python jupyter notebook. The predictive results revealed the potential of PCA-BPNN, which exhibited a high level of accuracy in comparison to the above models for all the considered variables. Two different model combinations were built for each single (RF, DT, KNN, SVM, AND BPNN) model and PCA algorithms (PCA-BPNN, PCA-DT, PCA-RF, PCA-SVM, and PCA-KNN). The results also depicted five models that demonstrated classification skill and therefore, can serve as reliable models. The classification is performed through PCA-BPNN (principal component analysis - back propagation neural network), the neuron network consists of nine (9) input vectors and 9 neurons in its output layer to recommend crops.77% accuracy is achieved when integrated principal component analysis with a back-propagating neural network (PCA-BPNN) is used. I conclude by saying that the outcomes of this thesis may contribute to the above-mentioned modeling of the treated soil parameters and provides a reference benchmark for solving the traditional way of farming in Ethiopia.