Exact solutions have always played and still play an important role in properly understanding the qualitative features of many phenomena and processes in various fields of natural science. Exact solutions of nonlinear equations, including those without a clear physical sense which do not correspond to real phenomena and pro- cesses, play an important role of test problems for verifying the correctness and assessment of accuracy of various numerical, asymptotic, and approximate methods. Moreover, the model equations admitting exact solutions serve as the basis for the development of new numerical, asymptotic, and approximate methods, which, in turn, enable us to study more complicated problems having no analytical solutions [1] . In the paper [2] , Laith and Nama introduced a new approach for solving second order linear differential equation with variable coefficients

To look for exact solution of (1) the authors introduced the substitution

and have looked for a solution of the Riccati equation

In this paper, we generalize the idea of [2] and propose a general approach for solving the nonlinear second order equation

which can be written as

where

In this section, we propose an algorithm that enables us to reduce the Equations (4) and (5) by looking for solutions of the partial differential equations

Theorem 1. If v(x; y) is any solution of (6) where (x; y) is a solution of (7), then Equation (4) can be reduced to a first order equation.

Proof. In order to prove this theorem, consider the transformation

if we differentiate both sides of (8) with respect to x we obtain

substituting (4) and (8) in (9), we have

assuming that is a solution of (7), Equation (10) can be reduced to (6), solving (6) for we have the result. ■

Theorem 2. If is any solution of the equation

Then (5) can be reduced to a first equation.

Proof. From theorem (1) the associated equation with is

which has a solution, thus the equation associated with is (11), solving for Equation (5) reduced to a first order equation. ■

Theorem 3. If is any solution of the equation

Then Equation (5) can be reduced to first order equation.

Proof. Equation (5) can be written as

applying theorem (1), we have that is a solution of

solving (13) for, the result follows. ■

Theorem 4. If, then Equation (5) can be reduced to a first order equation.

Proof. Applying theorem (2) the result follows. ■

In this section, we give some examples on our approach for reduction and finding solutions of nonlinear second order ordinary differential equations, these equations and more equations that can be easily solved by this method can be found in [1] [3] - [7] .

Example 1. Consider the equation

comparing with Equation (4) we note that, ,.

First, we solve

the associated ratios with Equation (17) are

from which, we find that

Second, we solve

the associated ratios with Equation (19) are

from which, we find that

Finally, we substitute, in Equation (8) to get

Example 2. Consider the equation

this equation can be written as

comparing with Equation (5) we have that, ,.

The equation associated with is

from which we find that. The equation associated with is

we look for a solution of the form

substituting in Equation (25), we have

Thus, and must satisfy the following equations

from which we find that

so,. Finally we solve

and two cases are considered,

, the solution is (34)

, the solution is (35)

Example 3. Consider the equation

Equation (36) can be written as

Comparing with Equation (5) we have, , furthermore

. So, theorem (4) can be applied as follows:

which implies that

Differentiating both sides of (39), we have

Assuming that, yields

thus, Equation (36) reduced to the first order exact ordinary differential equation

which has the solution

In this article, a new method is considered for solving second order nonlinear ordinary differential equations. The small size of computation in comparison with the computational size required by other analytical methods [1] , and the dependence on first order partial differential equations show that this method can be improved and introduces a significant improvement in solving this type of differential equations over existing methods. This method is proposed to be considered as an alternative approach being employed to a wide variety of equations.