I have to measure the S-parameters of an unequal dual-band power divider using VNA. The input impedence of the power divider is 50 ohm, but the output impedence of it is different than 50 ohm. is there anyway that I can set the impedences of the two ports of the VNA seperately during the calibration? if not, is there anyother way to test unequal-split power divider ? can anyone please help me about this?

You have an unequal power splitter, and you want the S-parameters of it. What will be the system impedance of the the other elemements connected to the power splitter. For example, does it come from a 50 ohm oscillator, and one arm goes to a 50 ohm cable to an amplifier, also 50 ohms, and the other arm goes to a 50 ohm cable to a detector, also 50 ohms....in this case, alll your impedances are 50 ohms, and will make your measurments relative to 50 ohms, and this implies that your splitter will be mismatched, if it is not lossy.

On the other hand, if your splitter is 50 ohms on the input, and one output is 75 ohms, and the other output is 50 ohms, and you want the S-parameters normalzed to each port differently (this is a very unusual situation,so I don't expect this is the case), then you must calibrate all 3 ports in one impedance, say 50 ohms, then use fixture simulator to transform the impedance of 1 port. AND, when you calibrate the 75 ohm port, you will calibrate it in 50 ohms but use an open, short and load where the load is an "arbitrary impedance" set to 75 ohms, so you can use your precision standards. Let me re-iterate that this is an unusual situation, suitable only for things like measuring transformers or minimum loss pads.