Almost everything commonly said about religion in Arabia before Islam comes from Muslim authors writing afterwards, and they were describing something their own tradition had superseded and named as ignorance.
That does not make the accounts worthless. The Book of Idols and the historical and exegetical literature preserve names of deities, the locations of their shrines, the tribes attached to them and the practices associated with pilgrimage, and much of it is corroborated where independent evidence exists. But it is testimony from the other side of a decisive change, gathered generations later, and organised around explaining what the new revelation replaced. Deities appear in it mainly to be refuted.
The independent evidence is growing and it comes from inscriptions. South Arabian kingdoms wrote extensively in their own scripts, and thousands of shorter texts in Safaitic, Thamudic and related scripts survive across the northern deserts, many of them prayers or invocations scratched by herdsmen. These give names and formulas that no later author had reason to invent, and they are steadily changing the picture.
Archaeology has been uneven for reasons of access and of politics, and some of the most important sites are in places where excavation has been difficult or where religious sensitivity limits what may be dug.
The honest summary is that the region's pre-Islamic religion is known in outline, that the outline comes largely from a hostile witness, and that the inscriptions are the check on it.