Internet of Things (IoT) solutions adopted by various time-sensitive application domains consist of fine-grained sensors to predict actionable knowledge more accurately. This paper presents the concept of IoT climate and IoT climate-based predictive models for future farming. Firstly, we consider an online agricultural forecasting service based on IoT data collected from weather stations in real-time. To demonstrate the effectiveness of our proposed platform, we designed a frost forecasting service and pest forecasting services on the IoT climate data collected from weather stations. Our system notifies the possibilities of frost for farmers so that they can protect crops against frost damage. And it also sends several pest forecast messages to farmers using push services. The proposed model can effectively provide more precise agricultural forecasting services thus could potentially not only improve crop yield but also reduce unnecessary costs, such as the use of non-essential pesticides and fertilizers.