A divider is used to (as the name suggests) reduce a higher voltage to a lower voltage. Using a divider on a small voltage will only make it smaller and harder to measure.
To get the most out of your measuring you need to have the reference voltage as close to the highest voltage you want to measure as possible. The highest voltage that can be applied is not dependent on the reference voltage, but the supply voltage, and as 1.2V is way less than 5V (or 3.3V on a 3.3V board) you have no worries there.
So you need to change the reference voltage - the voltage that the ADC uses as its upper measurement voltage. That's simple enough, since many of the Arduino boards include an internal 1.1V analog reference voltage generator that you can switch to (see: analogReference())
So when you have switched to the 1.1V analog reference the ADC is capable of measuring 0-1.1V in 1024 steps. That's 1.1/1023=0.00107V per bit of precision.
The formula is:
V = ADC / 1023 * Vref
So a reading of 1 would give you 0.00107V - a reading of 2 would be 0.00215V, etc. That's plenty of resolution to meet your 2 decimal places requirement.
One thing to note though is that the internal 1.1V reference isn't that stable or accurate (or not even there on some chips). If you need better stability and accuracy then it could pay to use an external voltage reference chip. These are like a linear voltage regulator but are far more accurate and stable (and can't provide lots of current). Pick one that is greater than your upper measurement voltage but as close to it as you can get (so just above 0.8V) and feed the output into the external ARef pin and switch the analog reference to the
You could use a voltage divider on the external ARef pin to provide a low reference voltage, but the stability and accuracy will be no better (and probably worse) than the internal 1.1V voltage reference, so there really is no point.