purposes. Obtaining crop-type maps early in the current growing season is highly advantageous for agricultural decisionmaking and management. Due to the abundance of high-resolution remote sensing data with precise spatial and temporal details, the frequency of data collection from various sources is anticipated to rise. The augmentation of data gathering will yield a larger volume of information available in the initial phases of the season. This study proposed an Adaptive Convolutional Neural Network with Incremental Training (ACNN-IT) method based on the CNN model. ACNN-IT is primarily intended to explore the potential of integrating diverse data sources for real-time crop-type identification. Given the periodic production of Sentinel remote sensing data, an adaptive strategy for early crop-type detection is necessary. Consequently, the incremental training-based CNN model is developed for autonomous feature extraction and classification. The proposed model comprises pre-processing, feature extraction, and classification. An iterative training method is utilized to familiarize the network with adaptive sentiment data and ascertain the ideal detection rate for each crop variety throughout the early growth season. An examination of the ACNN-IT model is conducted using remote sensing data collected from Sentinel-1 A and Sentinel-2 satellites. ACNN-IT strategy was validated using support vector machine (SVM), random forest (RF), and SoftMax classifiers. Simulation results revealed that the proposed model has improved the overall crop type detection accuracy by 3.56% with reduced computational burden by 7.23% compared to the state-of-the-art.