For resource reutilization, scrap tyres have long been investigated as an additive to concrete to form 'Rubcrete' for various applications and have shown promising results. However, the addition of rubber particles leads to the degradation of physical properties, particularly, the compressive strength of the concrete. In this study, a theoretical model was proposed to shed light on the mechanisms of decrease in compressive strength due to the addition of rubber particles as well as improvement in compressive strength through modification of particle surfaces. The literature suggests that the compressive strength can be improved by soaking the rubber particles in alkaline solution first to increase the inter-phase bonding between the rubber particles and cement. Instead, we discovered that the loss in compressive strength was due to local imperfections in the hydration of cement, induced by the addition of heterogeneous and hydrophobic rubber particles. Microscopic studies showed that the rubber particles disturbed the water transfer to create channels, which were prone to cracking and led to a loss in the compressive strength. Unexpectedly, no cracking was found along the surfaces of the rubber particles, indicating that the bonding strength between the rubber particles and cement phases was not the critical factor in determining the compressive strength. Therefore, a theoretical model was proposed to describe the water transfer in the Rubcrete specimens to explain the experimental data. In the model, the local water available for hydration (Q) is: Q = -A(slv)/6piv, where Q, A(slv), and v are mass flow rate (kg s(-1)), Hamaker constant (J), and dynamic viscosity (m2 s(-1)), respectively. By maximizing the quantity Q and, in turn, the Hamaker constant A(slv), the compressive strength could be improved. The Hamaker constant A(slv) for water film on rubber particle surfaces was smaller than that for the hydrated cement particles; the water transfer rate was lower in the presence of rubber particles because the Hamaker constant A(slv) for water film on rubber particle surfaces was smaller than that on the hydrated cement particles. Thus, the compressive strength of Rubcrete could be improved by increasing the Hamaker constant of the system. This was achieved by increasing the refractive indices of the solids (n(s)). The refractive indices of materials increase with increases in functional groups, such as OH and SH on the surface. The model provided a possible mechanism for the efficacy of treating rubber particles with NaOH in improving the compressive strength. By using NaOH solution treatment, an oxygen-containing OH group was formed on the rubber surface to increase the Hamaker constant of the system, leading to higher compressive strength. Based on this mechanism, a novel method for modification of the rubber particles was also proposed. In this process, the rubber particles were partially oxidized with hot air/steam in a fluidized bed reactor to produce the hydrophilic groups on the surface of the particles. Preliminary results obtained so far are promising in accordance with the theory.