Accelerometers are prone to vibrations, which I would imaging a model rocket would be producing in abundance. Gyros on the other hand are less prone to vibration but will rezero themselves (there is probably a technical term, but I don't know it), meaning that after a gyro is tilted at a certain angle for a period of time, that angle becomes the new level mark. So you basically need both a two axis accelerometer and a two axis gyroscope, using both readings-the accelerometer to know where level really is and the gyro to know when the rocket has tilted.
Now we get into the other problem. G-forces. Hobby level devices like these are designed to handle only a certain level of pressure before they are useless and/or decide to break. The lower the g-force tolerance the more accurate the device is, but the less rapid of change it can handle and I am pretty certain a rocket taking off is going to be producing quite a few Gs. So you will need to first figure out the amount of thrust your rocket is going to be producing so you can determine what level of accelerometer you will need (5G is the highest I have seen at the hobby level).
You can read more on accelerometers here: http://www.societyofrobots.com/sensors_accelerometer.shtml
How big is the rocket overall (length, diameter and fin sizes)? That, combined with the force of thrust will be your determining factor for what you use to move the fins. Most likely some type of micro or mini servo, but that is a complete guess at this point.
For microcontrollers (mcu) parallax makes a line called BasicStamp, which oddly enough is programmed in basic, and can probably handle what you are trying to accomplish. I believe there is a Basic compiler for the pic mcus as well. Or you can take a crack at programming in C and a whole world of mcus open up to you.
Now if I hear on the news about a homemade guided missile being launched at any point and time, I will be very put out and will require large sums of money for any further assistance.