Biofilms are aggregates of bacteria growing in extracellular matrix and polymers [1], [2], [3]. With a unique swarming growth advantage coupled with protection by the extracellular matrix, biofilms can colonize growth on virtually any surface [4], [5], [6]. Extracellular polymeric substances are able to provide a strong barrier to bacteria growing within biofilms against the threat of hostile environments to bacterial survival [7], [8]. EPS molecules fill and shape the space between cells in biofilms, directly determine the environmental and survival conditions of cells, and provide mechanical stability to biofilms [9].

In previous studies, it was found that biofilms develop a variety of different morphologies under different growth conditions [10], [11], [12], [13], [14]. These different morphologies reflect changes in various physicochemical properties of biofilms [15], such as adhesion properties, which are important for the effective removal of harmful biofilms attached to the surface of objects [16].

Many researchers have experimentally studied different branched forms of biofilms [17], [18], [19], [20], [21], [22]. The method of the experiment is simple and intuitive, but we are currently unable to investigate the mechanism of biofilm morphogenesis at lower scales because of the limitations of instruments and equipment. A more common and popular research method is modeling simulation. This type of research can be divided into two from the modeling perspective. One is to treat bacteria as a continuous density, describing the change in bacterial density. For example, Tsyganov, Deborah Schwarcz, and others studied different morphologies of biofilms by building mathematical models [23], [24]. Among all the mathematical models, the reaction-diffusion model, which is keen on by most researchers [25], [26], [27], [28], is usually used as the basis model to study biofilms. The other is the discrete model. One of the most important types in discrete models is agent-based model. Agent-based model can be used to study the relationship between agent and agent, as well as the mutual influence between agent and environment [29], [30], [31], [32]. It builds simulation models from the bottom layer, focuses on each agent's properties and the interaction between agents. Agent-based model has been applied by numerous researchers in recent years in the study of biofilms [33], [34]. Based on the characteristics of this model, Dmitri Volfson, Winkle, and William studied the sorting mechanisms of bacterial cells within biofilms and the spatial distribution of cells in different shapes [35], [36], [37], [38]. Although these studies are important to understand the mechanism of biofilm self-organization formation, little has been mentioned about the development of biofilm morphological structures.

In this paper, an agent-based particle model is proposed with NetLogo software, which models bacterial clusters into particle and defines the repulsive interactions among particles. The nutrient concentration of the substrate is defined as a continuous diffusion equation at the same time. Bacterial clusters multiply by taking up nutrients from the area (patch) on which they reside to divide, enabling sustained outward expansion. We investigate the effects of initial nutrient concentration and nutrient diffusion rate on biofilm morphogenesis and contrast with experiments, and both obtain good matches. In parallel, we introduce a second particle in the model to discuss the effect of interactions between bacterial cells and extracellular polymeric substances (EPS) on biofilm morphogenesis, and to our knowledge, this has not been reported.