A hyperspectral deep learning attention model was developed to predict the total chlorophyll content of greenhouse lettuce directly from hyperspectral images. The model uses a one-dimensional convolutional neural network (CNN) with a spectral attention module to estimate chlorophyll levels. Experimental results showed that this model outperformed traditional methods like partial least squares regression (PLSR) and random forest (RF), achieving an average R² of 0.746 and an average RMSE of 2.018. The spectral attention module helps the model learn the importance of different hyperspectral bands, improving its performance. The model was tested using hyperspectral data from 478 lettuce samples, with results validated through five-fold cross-validation. The study highlights the potential of deep learning and hyperspectral imaging for non-destructive, efficient chlorophyll estimation in leafy vegetables. The model's ability to process full hyperspectral data without additional wavelength selection or dimensionality reduction makes it suitable for real-world applications. The research also emphasizes the importance of developing methods for estimating chlorophyll content using deep learning to address the challenges of high-dimensional hyperspectral data. The findings suggest that the proposed model can be applied to various leafy vegetable species and different environmental conditions, offering a promising solution for automated monitoring and production management of leafy vegetables.A hyperspectral deep learning attention model was developed to predict the total chlorophyll content of greenhouse lettuce directly from hyperspectral images. The model uses a one-dimensional convolutional neural network (CNN) with a spectral attention module to estimate chlorophyll levels. Experimental results showed that this model outperformed traditional methods like partial least squares regression (PLSR) and random forest (RF), achieving an average R² of 0.746 and an average RMSE of 2.018. The spectral attention module helps the model learn the importance of different hyperspectral bands, improving its performance. The model was tested using hyperspectral data from 478 lettuce samples, with results validated through five-fold cross-validation. The study highlights the potential of deep learning and hyperspectral imaging for non-destructive, efficient chlorophyll estimation in leafy vegetables. The model's ability to process full hyperspectral data without additional wavelength selection or dimensionality reduction makes it suitable for real-world applications. The research also emphasizes the importance of developing methods for estimating chlorophyll content using deep learning to address the challenges of high-dimensional hyperspectral data. The findings suggest that the proposed model can be applied to various leafy vegetable species and different environmental conditions, offering a promising solution for automated monitoring and production management of leafy vegetables.