The most common approach to variation of Cp is a polynomial fit (keep the order low, not more than cubic), or for small ranges a simple linear fit is commonly used. Remember that extrapolation is always risky with curve fits. There are more complicated equations of state that are used in high precision work and extrapolate better, but these are rarely needed in practical problems.
Are you looking to determine the change in the value of heat capacity with temperature? If so, and you have measured Cp at only two temperatures, then these two data points (of Cp vs T), then a curve fit can have at most two degrees of freedom; i.e. two curve-fitting parameters. You need to have an idea of the appropriate functional form to use for curve-fitting, especially if you have only two data points. If you try to fit with a function with 3 or more fitting parameters; then the values of these parameters will not be uniquely determined using just two data points.
Enthalpy values are available for many fluids in Perry, Chemical Engineers Handbook. This also gives methods for calculating the behaviour of mixtures: not easy!. If your fluids are not conventional (i.e. as tabulated), measurements would be needed, or information from the supplier if they can provide it.