A multi-agent (MA) cellular automaton (CA) model framework for simulating grapevine growth and crop in Chardonnay cultivated in northern New Zealand is presented. Estimating or projecting grape crop (quantity of grapes in tons per hectare (ha) and berry quality in Brix (sugar content) is an extremely complex and challenging task as the crop depends on many factors that interact with each other at varying degrees and over different time intervals in a “chaotic” manner. These key factors and their influences are simulated using CA rules, MA behaviour and interactions. Two sets of CA lattices and rules are used to simulate individual grapevine growth and vineyard phonological dynamics. The results achieved show potential for simulating vine growth and yield in different grape varieties (Pinot Noir, Pinot Gris, Merlot and other wine styles) and scales, such as New Zealand’s major wine regions and that of world’s, in ways which that have not been explored previously.