Increased demand of farm products and depletion of natural resources compel the agriculture community to increase the use of Information and Communication Technology (ICT) in various farming processes. Agricultural Decision Support Systems (DSS) proved useful in this regard. The majority of available Agricultural DSSs are either crop or task specific. Less emphasis has been placed on the development of comprehensive DSS, which are non-specific regarding crops or farming processes. The crop or task specific DSSs are mainly developed with rule based or knowledge transfer based approaches. The DSSs based on these methodologies lack the ability for scaling up and generalization. The Knowledge engineering modeling approach is more suitable for the development of large and generalized DSS. Unfortunately the model based knowledge engineering approach is not much exploited for the development of Agricultural DSS. CommonKADS is one of the popular modeling frameworks used for the development of Knowledge Based System (KBS). The paper presents the organization, agent, task, communication, knowledge and design models based on the CommonKADS approach for the development of scalable Agricultural DSS. A specific web based DSS application is used for demonstrating the multi agent CommonKADS modeling approach. The system offers decision support for irrigation scheduling and weather based disease forecasting for the popular crops of India. The proposed framework along with the required expert knowledge, provides a platform on which the larger DSS can be built for any crop at a given location.