The computer simulation on self-assembling growth of aggregates of nanoclusters dispersed in randomly medium of colloidal solution is suggested. The Cluster-Cluster Aggregation model for original research self-organizing aggregations of nanoparticles is used. For description of random dynamics of nanoparticles dispersed in bounded volume, their aggregation is located inside a cubic cage of physical space with base length n = 30 composed from 27000 (30×30×30) equal-sized cubic cells. Aggregates of cubic-form nanoparticles are synthesized at different concentrations (10% - 100%). Time of evolution is dimensioned with discrete steps. On the first time step cubic particles occupy the cage of physical space randomly. On every time step particles migrate in one of the twenty six sectors. Every particle has nearest-neighbor vicinity, consisted from twenty six cells. As long as other particles come in this vicinity, the aggregate will be generated from these particles. Aggregates migrate in one of twenty six sectors just as well particles. If another particles or formed aggregates come in vicinity of this aggregate, it will cause new aggregate generation from these particles and/or aggregates. Process is to be continued, unless one cluster will include all particles generations. Analyzing topological graph of aggregates, in which vertexes denote particles, whereas ribs of graph denote chemical-physical bonds of neighboring particles, it is shown that increasing of number of aggregate graph ribs decreases enthalpy. That is a main gathered factor. The novelty of the exploration is a model validation that a maximum of the Shannon’s information of the aggregate topological graph reflects some self-assembling trends. When numbers of ribs countering each vertex increase from zero to a possible value it is revealed that the extreme property of Shannon’s information functional dependence on the concentration of nanoparticles reflects main morphologic features of the agglomeration.