No quantitative methods are currently available to measure different aggregation parameters in cell cultures. In this paper we describe a computer-based technique for the automatic and reliable analysis of cellular aggregates, starting from optical microscopy images of living cells grown in suspension. The method allows determination, on the same sample at different time intervals, of quantitative parameters, including aggregation percentage, average number of cells in aggregates, and aggregate size statistical distribution. To determine the number of cells in an aggregate starting from its two-dimensional microscopic profile, a model has been proposed and verified, using sphere packing theory. Algorithms have been tested on chondrocyte suspension cultures, where cell aggregation is a very early and critical event leading to cell differentiation. Using this technique for the analysis of chick embryo chondrocyte cultures, we observed that aggregate size and development kinetics depend on the culture conditions used. The method, with minor adaptations, is of potential use also in other cell systems to evaluate aggregation indexes or to study aggregation kinetics.