In the field of agriculture, choosing a right crop for a particular land is very important. For such recommendation, we need soil nutrients data and environmental data. Also, predicting the post-harvest price of the crop at the preparation stage helps the farmer to harvest the crop, which gives him profit. To recommend the three crops which can be cultivated on the land, Stacking Classifier with base models as K-Nearest Neighbors, Naive Bayes, Decision Tree and Support Vector Machine is used. The stacking classifier model is a machine learning ensemble which combines multiple base models to enhance accuracy. For these recommended three crops, the prices are predicted using LSTM model. The LSTM model is particularly useful in time-series forecasting. The accuracy of the crop recommendation model is 96.61% and Mean Absolute Percentage Error for the price prediction model is 5.6%. The system has the potential to increase agricultural production and profitability.