The problems of the digitalization of livestock enterprises are closely related to the construction of models and algorithms describing the functioning of individual technological processes and subsystems united by a common control system. Based on the cluster approach, three groups of tasks for the intellectualization and digitalization of objects in livestock breeding are formulated: 1) recognition of images of biological objects and models of their group and individual behavior, 2) genomic assessment of farm animals, prediction of their genetic potential, with the possibility of better adaptation to technologies and specifi c economic conditions, 3) multi-agent management of automated and robotic technical means. The authors initialized the video images of biological objects, developed a structural and functional model of a complex biotechnical system “Man-Machine-Animal”, including automated workstations of key specialists, signal receiving-and-transmitting base stations, technological modules for animal service (feeding, watering, milking, microclimate, etc.), representing local biotechnical systems. The paper presents a structural-and-logistic “funnel” model of a livestock farm functioning. The model includes vectors of incoming material fl ows, outgoing production fl ows and outgoing byproducts (production waste) described using appropriate formalizations. The authors provide the structural typifi cation of technological modules and subsystems for their mathematical analysis and subsequent digital transformation of livestock farms.