One of the most widely studied abilities in lizards is discrimination learning. The protocols used to test lizards are often novel or adapted from other taxa without proper validation. We need to ensure that tests of discrimination learning are appropriate and properly applied in lizards so that robust inferences can be made about cognitive ability. Here, we present a new protocol for testing lizard discrimination learning that incorporates a target training procedure, uses many daily trials for efficiency and reinforcement, and has a robust, validated, learning criterion. We trained lizards to touch a cue card using operant conditioning and tested lizards separately on a colour, and pattern discrimination test. Lizards successfully learnt to touch a cue card and to discriminate between light and dark blue but had issues discriminating the patterns. After modifying the test procedure, some lizards reached criterion, revealing possible issues with stimulus processing and interference of generalisation. Here, we describe a protocol for operant conditioning and two-choice discrimination learning in lizards with a robust learning criterion that can help researcher better design future studies on discrimination learning in lizards.