Rare-earth monosilicates (RE2SiO5) have been a promising material for the environmental barrier coating (EBC) owing to their superior thermal and mechanical properties: thermal conductivity from 1.87 W·m−1·K−1 for Dy2SiO5 to 3.48 W·m−1·K−1 for Y2SiO5 at room temperature; Young’s modulus from 144 GPa for Tb2SiO5 to 172 GPa for Lu2SiO5; flexural strength from 153 ± 3 MPa for Dy2SiO5 to 236 ± 10 MPa for Ho2SiO5. But there are still gaps in coefficient of thermal expansion (CTE) from various RE2SiO5 (6-10 × 10−6 K−1) and the matrix material (SiCf/SiCm, 4.5–5.5 × 10−6 K−1), which could lead to catastrophic coating failure. Exploring novel composition design in RE monosilicates is a critical challenge for next-generation EBCs. Entropy design in RE monosilicates is attracting considerable interest due to a combination of enhanced properties in EBC applications; however, trial and error strategies for various multi-entropy RE monosilicates are usually adopted based on empirical selection, which would take months, even years, for development cycles. Machine learning (ML) applied to next-generation EBC exploration can help accelerate the multi-entropy design in rare earth (RE) monosilicates. In this study, we propose an ML method, using lattice features of a series of multi-entropy RE monosilicates for predicting their lattice distortion and expansion behaviour at evaluated temperatures. The lattice structure for RE2SiO5 consists of the arrangement of [SiO4] tetrahedrons, and [REO6] and [REO7] polyhedrons. For simplicity, we alternatively use [SiO4] and [ORE4] tetrahedrons as structural units in the lattice of RE2SiO5. The distortion of [SiO4] and [ORE4] is described by the deviation of bond length and angle from regular tetrahedrons of Yb2SiO5. The approach is employed to predict the lattice distortion of over 20 new compositions. The suitability of the ML method here is demonstrated by comparing it with experimentally accurate lattice data from XRD refinement of 8 multi-entropy RE monosilicates. The synthesizability of 8 multi-entropy RE monosilicates is investigated by means of simultaneous measurement of weight change (TGA) and true differential heat flow (DSC). Incorporating the multi-entropy design into the RE monosilicates provides an opportunity for tuning the intrinsically physical properties and their further potential to predict their thermal expansion behaviour.