Correlation matrix is usually a table giving you all associations between variables you are applying in your analysis. Also you are provided with p values indicating if these associations are statistically significant. Based on those correlations you may develop regression models, or some more complicated ones as SEM. And what is important correlation does NOT imply causation.
It's simply a tabular (square, symmetrical) arrangement of correlation coefficients such that the variables are listed as rows and again as columns and the intersection of a given row with a given column provides the correlation coefficient between these two variables. Thus, if one has three variables X1, X2, & X3, the correlation matrix would be a 3x3 matrix and the entry 2,3 would give the correlation coefficient between the 2nd and 3rd variables. It is used in the development of multivariate methods like factor analysis because it provides a systematic measure of the relationship among an arbitrarily large number of variables.