Abstract BACKGROUND Kochia [ Bassia scoparia (L.) A.J. Scott] and waterhemp [ Amaranthus tuberculatus (Moq.) J.D. Sauer] are among the most aggressive and competitive weed species in sugarbeet production. Their similarity to the crop during early growth stages poses a significant challenge for early identification using conventional imaging techniques. This study aimed to develop and evaluate a hyperspectral imaging based deep learning model capable of distinguishing kochia and waterhemp from sugarbeet under field conditions. Hyperspectral images were acquired and preprocessed to extract spectral and spatial information for classification. RESULTS The attention enhanced convolutional neural network (AE‐CNN), trained using the combined spectral and spatial features, achieved the highest performance with a classification accuracy of 99.99%, and precision, recall, and F1‐score values of 1.0. In comparison, the support vector machine (SVM), trained using only spectral features achieved a classification accuracy of 96.98%, with precision, recall, and F1‐score values of 0.97. CONCLUSION These results highlight the potential of ground‐based hyperspectral imaging to accurately distinguish invasive weed species from crops, supporting site‐specific weed management in agriculture. The findings contribute valuable insights into the utilization of plants spectral signatures for early‐stage weed identification and support the development of timely and targeted weed control strategies. © 2025 Society of Chemical Industry.