The Internet of Things (IoT) has quickly developed in nowadays and enormously influenced in numerous industries. Therefore, IoT technology has enhanced productivity in numerous sectors and smart farming has hugely helpful from IoT. The smart farming allows agriculture, high crop yield and effective usage of natural source to maintain long time. This research proposed a Kernel K-means with Support Vector Machine (SVM) for smart agriculture system based on IoT. The collected dataset is used here and it is preprocessed by data cleaning and classified through Kernel K-means+SVM. The k-means efficiently cluster the data which is nonlinearly separable through changing data into high-dimensional space where it is linearly separable. The SVM is applied after k-means which achieve better classification performance. The initial clustering assist to segment data into groups which simplifies SVM to find optimal hyperplane for classification. The metrics such as accuracy, precision, recall, specificity and f1-score are used for evaluating Kernel K-means+SVM performance. It achieves 0.996 accuracy, 0.992 precision, 0.991 recall, 0.993 specificity and 0.994 f1-score when compared to K-means+SVM, Long Short-Term Memory (LSTM).