Scientists differ among themselves regarding the scientific status of multiverse theories. Some, such as George Ellis, don’t think multiverse theories are testable, and hence not scientific. Others, think multiverse models are (or could be) testable, and hence are scientific. Many Christian apologists have sided with Ellis et al and rejected the multiverse as a valid scientific theory on the grounds that it is not testable. Some, including myself, have argued that multiverse theories are not based on the evidence, but ad hoc theories invented by cosmologists to get around the theistic implications of fine-tuning in physics.
Jeff Zweerink from Reasons to Believe wrote a short article addressing the scientific nature of and foundation for multiverse theories. He argues that some multiverse models do make testable predictions (even if we are currently unable to test those predictions empirically), and thus should be “included in the realm of scientific investigation (while stopping short of taking a firm position on the demarcation question –whether multiverse theories qualify as scientific).
More importantly, he argues that at least some multiverse theories are based on other scientific findings, and not invented whole-cloth for the purpose of answering the fine-tuning problem: