Based on resource endowments, West African farmers can be classified into small scale, medium scale, or large scale. Each classification faces a decision-making challenge when choosing crops and allocating resources for maximum returns. This is more difficult due to frequent variabilities in weather, agro-input, and commodity prices. This paper employs the Decision Support System for Agro-technological Transfer (DSSAT) crop model to simulate the yields of six crops (maize, cowpea, soybean, rice, groundnut, and millet) in the Northern and the Upper East regions of Ghana for a period of 20 years (1990–2010) for each farmer classification. The simulated yields were input to a linear programming (LP) framework that optimized the resource allocation based on farmgate prices and production costs. We also employed the ENSO to explore the value of seasonal weather forecasting on farm plan optimization. Our results indicate that the weather variability has impact on crop yields as captured by the DSSAT model. The optimized resource allocation differed year to year, suggesting an annual update of the optimized farm plan for each farmer type. Combining a dynamic crop and economic optimization models provides a good basis for enhancing agricultural planning and hence, should be promoted for use to support farmers decision-making.