Tomato leaf diseases have a significant impact on tomato cultivation modernization. However, traditional diagnostic approaches suffer from low efficiency, misclassification, and inability to adapt to complex field environments. Additionally, existing models struggle with intra-class variability and inter-class similarity, reducing their reliability in real-world disease management. This research aims to address these challenges by introducing a novel TOMTO-NET approach for accurate tomato leaf disease detection using Dual-MoNet. A Gaussian Star Filter (GaSF) is employed to reduce noise while preserving essential disease features in tomato leaf images. A MobileNet backbone integrated with a Dual Attention Block (Dual-MoNet) is used for efficient feature extraction, where channel-wise and spatial attention mechanisms enhance fine-grained disease representation. A Spiking Neural Network (SNN) is then utilized for biologically inspired classification of tomato leaves into Healthy and Diseased categories. The effectiveness of the TOMTO-NET approach was evaluated using F1 score, recall, specificity, accuracy, and precision. The experimental results demonstrate that the proposed TOMTO-NET approach achieves an overall accuracy of 98.98%. The proposed TOMTO-NET method improves the accuracy by 1.15%, 2.55%, and 4.15% compared to DM-YOLO, ToLeD, and PLPNet, respectively