Scientific Model
A scientific model is a simplified representation of a system, such as a diagram, equation, or physical replica, used to explain and predict.
Definition
A scientific model is a simpler version of something real that helps you understand it: a drawing, a small copy, a computer program, or a math rule. It represents an object, system, or process by capturing the important features so they can be studied, explained, and used to make predictions, and models can be physical (a ramp and cart standing in for a hillside), conceptual (a particle diagram), mathematical (an equation), or computational (a simulation). Every model encodes a hypothesis about how the system works and rests on idealizations such as frictionless surfaces or point masses that define where it applies. Its value is judged by explanatory power, predictive accuracy within its domain, and simplicity, and systematic disagreement between predictions and careful data is how models are refined or replaced.
Example
A drawing of salt breaking into tiny pieces that spread out between water particles is a model of dissolving: you cannot see the particles, but the picture explains why the salt disappears. The particle model predicts that warmer water dissolves salt faster because faster-moving particles collide with the crystal more often, and measured times of $48\ \text{s}$ at $10^\circ\text{C}$ and $14\ \text{s}$ at $50^\circ\text{C}$ agree. The model $v = \sqrt{2gh}$ for a cart released from height $h$ ignores friction and wheel rotation; a prediction of $2.8\ \text{m/s}$ against a measured $2.6\ \text{m/s}$ shows it runs about $7\%$ high, and adding a rotational term and a friction estimate closes most of the gap.
Key Insight
A model leaves things out on purpose; a globe does not show every house, and that is what makes it useful. Models are tools, not truths, and the same phenomenon can have several useful models depending on the question being asked. All models are wrong in some respect, and the useful ones are wrong in ways you know about, which is why stating a model's assumptions matters as much as stating its equation.