Context : The mass distribution in both galaxy clusters and groups is an important cosmological probe .
It has become clear in the last years that mass profiles are best recovered when combining complementary probes of the gravitational potential .
Strong lensing ( SL ) is very accurate in the inner regions , but other probes are required to constrain the mass distribution in the outer regions , such as weak lensing or dynamics studies .

Aims : We constrain the mass distribution of a cluster showing gravitational arcs by combining a strong lensing method with a dynamical method using the velocities of its 24 member galaxies .

Methods : We present a new framework were we simultaneously fit SL and dynamical data .
The SL analysis is based on the LENSTOOL software , and the dynamical analysis uses the MAMPOSS t code , which we have integrated into LENSTOOL .
After describing the implementation of this new tool , we apply it on the galaxy group SL2S J02140-0535 ( z _ { spec } = 0.44 ) , which we have already studied in the past .
We use new VLT/FORS2 spectroscopy of multiple images and group members , as well as shallow X-ray data from XMM .

Results : We confirm that the observed lensing features in SL2S J02140-0535 belong to different background sources .
One of this sources is located at z _ { spec } = 1.017 \pm 0.001 , whereas the other source is located at z _ { spec } = 1.628 \pm 0.001 .
With the analysis of our new and our previously reported spectroscopic data , we find 24 secure members for SL2S J02140-0535 .
Both data sets are well reproduced by a single NFW mass profile : the dark matter halo coincides with the peak of the light distribution , with scale radius , concentration , and mass equal to r _ { s } = 82 ^ { +44 } _ { -17 } kpc , c _ { 200 } = 10.0 ^ { +1.7 } _ { -2.5 } , and M _ { 200 } = 1.0 ^ { +0.5 } _ { -0.2 } \times 10 ^ { 14 } M _ { \odot } respectively .
These parameters are better constrained when we fit simultaneously SL and dynamical information .
The mass contours of our best model agrees with the direction defined by the luminosity contours and the X-ray emission of SL2S J02140-0535 .
The simultaneous fit lowers the error in the mass estimate by 0.34 dex , when compared to the SL model , and in 0.15 dex when compared to the dynamical method .

Conclusions : The combination of SL and dynamics tools yields a more accurate probe of the mass profile of SL2S J02140-0535 up to r _ { 200 } .
However , there is tension between the best elliptical SL model and the best spherical dynamical model .
The similarities in shape and alignment of the centroids of the total mass , light , and intracluster gas distributions add to the picture of a non disturbed system .