Capital budgeting decisions faced by tart cherry producers often challenge our traditional valuation techniques. Real Options Valuation (ROV) methods may be useful but assumptions of existing ROV approaches are restrictive and, in some cases, unrealistic. In this paper we assert that use of existing option pricing methods can not be justified. Instead, dynamic programming approach is more appropriate. We develop a multi-period model and use it to obtain an optimal orchard replacement policy. The model is applied to an example farm from Northwestern Michigan and the results provide the following messages. First, flexibility options can be estimated for individual tart cherry producers using the DP approach albeit, indirectly. Second, a farmer who uses the DP approach to develop contingency optimal replacement rules will be better off than one who uses an ad hoc standard replacement rule. Third, if the SW climate scenario shifts to NW Michigan, tart cherry orchard values may fall substantially with implications on the future of tart cherry production in that region, unless compensating price increases follow.