Agriculture has a strategic role in a country whereas food self-sufficiency being the main goal to be achieved. Indonesia has set a strategic plan for increasing the productivity of several commodities, including rice, especially irrigated lowland rice. That matter can be done by agricultural land extensification, which requires a land suitability directional map. This study aims to produce irrigated lowland rice land suitability maps which can be obtained by evaluation using spatial decision tree algorithm. The model is made in two different types, where model Y is an optimized version of model X. The dataset consists of two categories, namely eleven explanatory layers which are land and weather characteristics, and a target layer that represents irrigated lowland rice land suitability in study area of Grobogan Regency, Central Java Province. As an addition to planting requirements, two spatial weather datasets were generated using ordinary cokriging interpolation, which was not used in previous research, while actually being important element for determining plant timing an agricultural commodity. Based on accuracy, model Y is the best model with 96.67%, compared to model X with 86%. Both models make relief variable as the root node, but in spatial decision tree result, model X involves all variables, while model Y does not involve an elevation variable. The addition of weather variables in models is appropriate, as evidenced by the involvement in rules.