Try these:

Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2014). Multivariate Data Analysis. Prentice-Hall, Inc. London: Pearson Education. http://doi.org/10.1038/259433b0

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston: Pearson Education.

Thank you Dr. Umar. I will explore it. 

In agriculture, nonlinear regression has been used to model crop response to some soil attribute or continuous treatment (such as crop yield vs. soil P, or crop yield vs. applied N).

A few common models are the Mitscherlich-Bray, linear-plateau, and quadratic plateau.

These might be useful in agricultural extension to show that increasing inputs beyond some point result in diminishing returns in crop yield, or in no return at all.

Thank you Dr. Mangiafico for your valuable suggestions. Actually, I want to apply non-linear regression model taking  some socio-economic variables as causal variables with type of farmers queries in IVRS based agro-adviosry system separately taking as dependent variables.

As long as both variables are continuous, you can do this.

But you'll have to determine what model you'd like to fit.

You might also ask yourself what the goal is of fitting nonlinear models to your data.  If it's just to fit a pretty line, that's okay.  But often there is some parameter in the non-linear model that is of interest.  For example, in a linear plateau model, ( http://rcompanion.org/handbook/images/image148.png  ), the y-value of the plateau and the x-value where the segments join may be of interest.

I'll also mention that for data which don't fit curves well, a Cate-Nelson model is useful for breaking bi-variate data into groups of low-x-and-low-y and high-x-and-high-y ( http://rcompanion.org/rcompanion/h_02.html ).

Thank you Dr. Mangiafico for your guidance. I will explore all the possibilities..